Trade In Knowledge 9781108490429, 9781108780919, 9781108748476

Citation preview

TRADE IN KNOWLEDGE

Technological change has transformed the ways knowledge is developed and shared internationally. Accordingly, in the quarter-century since the WTO was established, and since its Agreement on Trade-Related Aspects of Intellectual Property Rights came into force, both the knowledge dimension of trade and the functioning of the IP system have been radically transformed. The need to understand and respond to this change has placed knowledge at the centre of policy debates about economic and social development. Recognizing the need for modern analytical tools to support policymakers and analysts, this publication draws together contributions from a diverse range of scholars and analysts. Together, they offer a fresh understanding of what it means to trade in knowledge in today’s technological and commercial environment. The publication offers insights into the prospects for knowledge-based development and ideas for updated systems of governance that promote the creation and sharing of the benefits of knowledge.   is Director of the Intellectual Property, Government Procurement and Competition Division of the WTO Secretariat.   is a former senior officer in the Intellectual Property, Government Procurement and Competition Division of the WTO Secretariat.

TRADE IN KNOWLEDGE Intellectual Property, Trade and Development in a Transformed Global Economy

Edited by ANTONY TAUBMAN JAYASHREE WATAL

University Printing House, Cambridge CB2 8BS, United Kingdom One Liberty Plaza, 20th Floor, New York, NY 10006, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia 314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre, New Delhi – 110025, India 103 Penang Road, #05–06/07, Visioncrest Commercial, Singapore 238467 Cambridge University Press is part of the University of Cambridge. It furthers the University’s mission by disseminating knowledge in the pursuit of education, learning, and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781108490429 DOI: 10.1017/9781108780919 © 2022 The World Trade Organization This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2022 Printed in the United Kingdom by TJ Books Limited, Padstow Cornwall A catalogue record for this publication is available from the British Library. ISBN 978-1-108-49042-9 Hardback ISBN 978-1-108-74847-6 Paperback Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

CONTENTS

List of Figures page ix List of Tables xii List of Contributors xiv Preface xxxi 1 Thematic Overview: Charting the Evolution of Knowledge Flows 1     

 

Conceptual Framework

25

2 The Shifting Contours of Trade in Knowledge: The New ‘Trade-Related Aspects’ of Intellectual Property 27  

3 How Digitization is Transforming Trade

84

 ,     

4 Intellectual Property and Digital Trade – Mapping International Regulatory Responses to Emerging Issues 108  . -   

 

Measuring Trade in Knowledge

169

5 Measuring International Intellectual Property Transactions in a Globalized World: Current Challenges and Possible Improvements 171     

v



vi

6 A Missing Link in the Analysis of Global Value Chains: Cross-Border Flows of Intangible Assets, Taxation and Related Measurement Implications 194     -

7 Global Ebbs and Flows of Patent Knowledge

218

 . ,  .    . 

8 Sources of Knowledge Flow between Developed and Developing Countries 265  ,      

9 Using Intellectual Property Data to Measure Cross-Border Knowledge Flows 293  ,  . ,  .    . 

10 The Global Digital Content Landscape

323

 . 

11 Cross-Border Knowledge Flows through R&D FDI: Implications for Low- and Middle-Income Countries

352

 ,  ,  ó   

12 The Innovation Environment and Knowledge Diffusion: Improving Policy Decisions through Patent Analytics 376  ,  .    

 

Impact of Knowledge Flows on Trade and Development 403

13 Global Knowledge Flows, Absorptive Capacity and Capability Acquisition: Old Ideas, Recent Evidence and New Approaches 405     . 

14 Trade in Intellectual Property-Intensive Goods     

431

 

vii

15 Knowledge Spillovers through International Supply Chains 453     ˊˊ

16 How do Patents Shape Global Value Chains? International and Domestic Patenting and Value-Added Trade 471  .    . 

17 The Enforcement of Intellectual Property Rights in a Digital Era 498  .    

18 The Digital Creative Economy and Blockchains: Options and Prospects for the Developing World 531  ,  ,     

 

Policy, Regulatory and Legislative Frameworks 553

19 Streaming of Music and Audiovisual Works

555

 

20 Adapting Trade Rules for the Age of Big Data

591

 

21 Trade in Knowledge and Cross-Border Data Flows: A Look at Emerging Digital Regulatory Issues 623  

22 Cross-Border Knowledge Flows under International Trade Agreements: A Need for New Multilateral Disciplines? 649      

23 The Need for a Global Framework for Knowledge Transactions: Cross-Border Licensing and Enforcement 685      . 

24 Fitting Machine-Generated Data into Trade Regulatory Holes 738  . 



viii

 

Looking Forward

769

25 Looking Forward: Building the Foundations for Policymaking in the Knowledge Economy 771  

Index

807

FIGURES

2.1 2.2 2.3 4.1 4.2 5.1 5.2

5.3 6.1 6.2 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 7.10 7.11

List of exporters in recorded optical discs in 2018 page 36 List of exporters in unrecorded optical discs in 2018 37 Value and share of trade in physical form of digitizable goods 53 RTAs with increased IP content over time 132 Specific IP provisions in RTAs with IP 135 World international transactions relating to charges for the use of intellectual property, n.i.e., 1995–2017 175 World international transactions relating to charges for the use of intellectual property n.i.e. (CUIP), computer services (CS), research & development (R&D) and audio-visual services (AV), 2005–2017 175 European Union (28) and the United States – charges for the use of intellectual property, n.i.e., receipts by main partner, 2016 176 Royalties Licences and Fees (RLF) receipts and payments, in USD billion and as share of trade 199 Charges for the use of IP as a percentage of lagged R&D, 2016 205 Trends in patent numbers 230 Rates of interjurisdictional citation 231 Patent citations between developed and developing countries 233 Importance of citing patents between developed and developing countries 234 Importance of cited patents between developed and developing countries 235 Mean importance of citing patents between developed and developing countries 236 Mean importance of cited patents between developed and developing countries 237 Trends in total, wind power and pharmaceutical patents over time 238 Trend in wind power patents over time 239 Rates of interjurisdictional citations for wind technology 240 Wind power patent citations between developed and developing countries 240

ix

x

  

7.12 Importance of citing wind power patents between developed and developing countries 241 7.13 Importance of cited wind power patents between developed and developing countries 242 7.14 Mean importance of citing wind technology patents between developed and developing countries 243 7.15 Mean importance of cited wind technology patents between developed and developing countries 244 7.16 Trend in pharmaceutical patents over time 245 7.17 Rates of interjurisdictional citation for pharmaceutical technology 246 7.18 Pharmaceutical patent citations between developed and developing countries 246 7.19 Importance of citing pharmaceutical patents between developed and developing countries 247 7.20 Importance of cited pharmaceutical patents between developed and developing countries 248 7.21 Mean importance of citing pharmaceutical patents between developed and developing countries 250 7.22 Mean importance of cited pharmaceutical patents between developed and developing countries 250 7.23 Radial diagram of international knowledge flow 252 7.24 Directional knowledge flow between developed and developing world countries 253 8.1 R&D collaboration (applicant criterion), by region and technology 271 8.2 R&D collaboration (inventor criterion), by region 274 8.3 Technology sourcing from developing countries 276 8.4 Technology sourcing from developed countries 277 8.5 Technology transfers to developing countries 279 8.6 Technology transfers from developing countries 281 A8.1 Links between the United States and Canada/Core Europe and developing countries 292 9.1 USPTO patent assignments at grant, 1976–2017 314 9.2 USPTO trademark assignments, 1899–2018 314 10.1 Breakout of the video on demand (VoD) ecosystem, by types of providers (non-exhaustive) 333 10.2 Global music revenue, by segment, 2017 338 11.1 Dynamics of R&D&DDT greenfield investments (total and %) 358 12.1 Percentage shares in top 10 per cent most valuable crop protection patents as measured by the Patent Asset Index 385 12.2 Process of concentration in the chemical industry 386

  

xi

12.3 Sum of Patent Asset Index (top) and average Competitive Impact (bottom): development from 2000 to 2015 for top 10 countries by Patent Asset Index in 2015 388 12.4 Top 10 countries where German patents spread in terms of the Patent Asset Index from 2000 to 2015 391 12.5 Top 10 countries where Japanese patents spread in terms of the Patent Asset Index from 2000 to 2015 392 12.6 Top 10 countries where US patents spread in terms of the Patent Asset Index from 2000 to 2015 393 15.1 R&D spillovers by distance and the intensity of supply chain linkages, the effect of a 10 per cent increase in foreign R&D spending 462 16.1 Construction of final merged dataset from PATSTAT and WIOD data 478 16.2 Yearly VAX, NStages and Upstreamness and international patent flows by sector 480 16.3 Estimated unilateral elasticities for VAX, NStages and Upstreamness with respect to patents with 95 per cent confidence intervals based on robust standard errors 486 16.4 Regression coefficients for VAX, NStages and Upstreamness based on patent flows by sector 488 16.5 Unilateral regression coefficients for VAX, NStages and Upstreamness based on patent flows by region 489 17.1 Bollywood revenue (log scale), 1960–2010 505 17.2 Indices of Indian movie production, 1960–2010 506 17.3 IMDb rating of top 20 Bollywood movies 507 17.4 Google notice and takedown requests, 2011–2016 513 17.5 Post-shutdown change in weekly digital movie sales vs December 2011 Megaupload penetration 515 17.6 iTunes single track unit sales trends (2008–2011) 519 18.1 The economic contribution of cultural and creative industries by region, USD billion 534 18.2 Global recorded music revenues, 2004–2019, USD billion 538 18.3 Global music revenues, 2019 539 18.4 Global collections of royalties, share by regions, 2017 541

TABLES

A3.1 Specific commitments across selected sectors, by modes of supply number of schedules with specific commitments (out of 152) page 105 4.1 Selected ISP provisions in selected EU and US RTAs 149 5.1 Exports of high-tech goods 1995, 2000, 2010, 2015–2017 174 5.2 Reporting of BPM5 intellectual property-related items, 2010 178 5.3 Reporting of BPM6/MSITS2010 intellectual property-related items, 2015 182 6.1 Taxonomy of formalized and informal technology transactions 197 6.2 Regression results of excess cross-border receipts to tax rates 206 6.3 Estimates of global and low-income country fiscal effects from profit-shifting 208 6.4 Estimate of global profit-shifting of cross-border CUIP flows 209 A8.1 Groups of countries used in the chapter 287 9.1 Trade in knowledge: summary of selected literature 297 10.1 Global digital content revenues, by sector and subsector, 2017 325 11.1 Geographical distribution of R&D&DDT FDI (total and %) 359 11.2 Geographical and sectoral distribution of R&D&DDT FDI (total and %) 360 11.3 R&D&DDT greenfield investments towards low- and middle-income countries: main host and home countries (total and %) 362 11.4 Global cities attracting R&D&DDT greenfield investments in low- and middle-income countries (total and %) 363 11.5 Top R&D&DDT investors in low- and middle-income countries (total FDI) 365 11.6 R&D-related acquisitions in low- and middle-income economies 366 12.1 Technological dependence measured via Patent Asset Index percentage changes between 2000 and 2015 in prior art from selected countries’ origination 394 14.1 List of economies by economy-type 438 14.2a Definition of high-IP group (SITC Rev. 3 Labels and Codes) 442 14.2b Definition of high-IP clusters (SITC Rev. 3 Codes) 443

xii

  

xiii

14.2c Definition of control group of low-IP products (SITC Rev. 3 Labels and Codes) 444 14.3 Descriptive statistics 445 14.4 Post-TRIPS estimated effect on trade of high-IP sectors, 1993–2016 446 A14.1 Classification criteria for a country’s TRIPS compliance 451 15.1 Summary of main variables by industry, average over the period 2000–2008 458 15.2 Summary of main variables by economy, average over the period 2000–2008 459 A15.1 Industry descriptions 467 A15.2 10 most connected economy-pair-industries by the type of supply chain linkages 468 A15.3 Knowledge spillovers through supply chains – regressions results 469 A15.4 Knowledge spillovers from main innovators – regression results 469 A15.5 The distance decay of knowledge spillovers 470 A16.1 Country-Industry Regressions on VAX Ratio, 1995–2011 496 A16.2 Country-Industry Regressions on NStages and Upstreamness, 1995–2011 497 17.1 Academy awards by country, pre-/post-piracy 508 17.2 Taxonomy of anti-piracy options 514 A17.1 Peer-reviewed journal articles finding no statistical impact of piracy 525 A17.2 Peer-reviewed journal articles finding that piracy harms sales 526 18.1 Music collections by region, 2017 (Euro millions) 540 18.2 Average collection for copyright, by number of inhabitants across the globe 541 A18.1 Digital music providers in the developing world 550 22.1 Provisions relating to knowledge flows in WTO agreements and selected FTAs 668

CONTRIBUTORS*

            is an Assistant Professor in Political Economy at the University of Bari. His research is focused on international trade and development economics. He has an MSc in Economics and Econometrics from the University of Essex, a PhD in Economics from the University of Bari and a PhD in Economics from the University of Glasgow. He has working experience with UNIDO, the London School of Economics and the University of Pavia.              is a Professor of Economics and Public Policy at Carnegie Mellon University, where he directs the Future of Work initiative at the Block Center for Technology and Society. He is also a nonresident senior fellow at the Peterson Institute for International Economics and a research associate of the National Bureau of Economic Research. He served as a Senior Economist at the Council of Economic Advisers in 2011–2012. Branstetter has conducted research in the domains of innovation, the economics of intellectual property, international technology transfer and international trade and investment.      is a Senior Lecturer at the Faculty of Law of the University of Lucerne, Switzerland. She teaches international intellectual property, media, and Internet and trade law. Mira’s current research interests are in the areas of digital trade, culture, copyright, data protection and Internet governance. Mira is the principal investigator of ‘The Governance of Big Data in Trade Agreements’ project, which is sponsored by the Swiss National Science Foundation. She consults for the European Parliament, UNESCO and others on issues of digital innovation and cultural diversity. Mira has co-edited the publications Trade Governance in the Digital Age (Cambridge University Press 2012) and Big Data and Global Trade Law (Cambridge University Press 2020). She is the author of Public Service * Correct as of 2020.

xiv

   

xv

Broadcasting 3.0: Legal Design for the Digital Present (Routledge 2015). Mira’s publications are available at: http://ssrn.com/author=483457.                 is Counsellor in the Trade in Services and Investment Division at the World Trade Organization (WTO). She is an economist with a BSc degree from Bocconi University, Italy, and an MA degree from the College of Europe, Belgium, and the London School of Economics, UK. She has worked in the WTO since 1998. She is the Secretary of the WTO Council for Trade in Services and deals, inter alia, with issues relating to the movement of natural persons (mode 4), various transport services and e-commerce. She has conducted many technical assistance activities focused on services trade and the disciplines of the General Agreement on Trade in Services, including for economies seeking to accede to the WTO. Ms Carzaniga has published extensively on matters relating to trade in services and has contributed to many expert fora, conferences and academic seminars.                is a full Professor in Innovation Studies at Lund University (Sweden) and Adjunct Professor at Aalborg University (Denmark). She holds a BA and PhD in Economics from the Autonomous University of Madrid (UAM). The focus of her research is on global networks and the processes of knowledge creation and adoption that underpin innovations and the transformations towards sustainable forms of living (economic, social and environmental). She has actively worked on innovation in developing countries such as China, India, South Africa, Thailand and Brazil for over 20 years. She has been an advisor to international organizations such as the European Commission, UNCTAD, OECD and UN-ECLAC. She has published in international journals, refereed books and handbooks in the fields of innovation, development studies and knowledge management such as research policy, industry and innovation, innovation and development and european planning studies. She is editor of the Journal of Innovation and Development, the International Journal of Innovation and the Journal of Sustainability Research.           is Assistant Professor at Telecom ParisTech, and Affiliated Research Fellow at the Max Planck Institute for Innovation and Competition (MPI). Prior to that, she was Senior Research Fellow at the MPI, affiliated with the Innovation and

xvi

    

Entrepreneurship research group. Laurie obtained her PhD in Economics from Mines ParisTech-PSL in 2017. She also holds a Masters degree in Economics of Markets and Organizations from the Toulouse School of Economics. In her research, Laurie conducts empirical studies in economics of innovation, with a focus on the determinants and consequences of markets for technology. She is particularly interested in the interplay of these markets and taxation, digital transformation and firm strategy. Laurie’s PhD dissertation received the Best Dissertation award, finalist of the TIM division of the Academy of Management 2018 in Chicago.       is an associate director at the Information Technology and Innovation Foundation covering trade policy at the Information Technology and Innovation Foundation. He focuses on cross-border data flows, data governance, intellectual property, and how they each relate to digital trade and the broader digital economy. He previously worked as a researcher in the Southeast Asia Program at the Center for Strategic and International Studies. Prior to that, he worked for eight years in Australia’s Department of Foreign Affairs and Trade, which included positions working on G20 global economic and trade issues and the Doha Development Round. Cory also had diplomatic postings in Malaysia, where he worked on bilateral and regional trade, economic and security issues, and in Afghanistan, where he was the deputy director of a joint US–Australia provincial reconstruction team. Cory holds an MA degree in public policy from Georgetown University and a BA degree in international business and commerce from Griffith University in Brisbane, Australia.                is Associate Professor of Strategy and Innovation at Copenhagen Business School and Research Scientist at the MIT Innovation Initiative. She serves as Senior Institute Associate at the Institute for Strategy and Competitiveness at Harvard Business School. Delgado’s research focuses on the relationship between the regional business environment and the performance of inventors, firms, regions and countries. She examines the role of regional clusters – geographic concentrations of related industries, firms and supporting institutions – in job creation, innovation, entrepreneurship, inclusivity and resilience. Delgado has developed new methods for defining and mapping industry clusters and the supply chain economy, providing

   

xvii

tools to help firms, practitioners and policymakers create regional and national strategies.     .   is the Director of the BBG Information Policy Institute, Houston, Texas. He is the author of Modern Licensing Law (Thomson West 2019) and a new edition of Drafting Effective Contracts (Wolters Kluwer 2020), both of which he previously co-authored with the late Professor Raymond T. Nimmer of the University of Houston Law Center. He and Lorin Brennan are updating Professor Nimmer’s Information Law and The Law of Computer Technology (both with Thomson West 1985) and are planning new editions for both. He has written and lectured extensively on licensing and information law topics, has been an adjunct at the University of Law Center and has been a lecturer at Beijing Normal University. He is currently a partner at Hunton Andrews Kurth, LLP.             was an extern with the Office of the Chief Economist at the United States Patent and Trademark Office (USPTO). He is currently a Research Economist in the Business Research Division at the University of Colorado, Boulder, where he focuses on researching and forecasting the Colorado and US economies. His research interests include international trade, economic policy and machine learning. Jacob holds an MA in Global Finance, Trade and Economic Integration from the University of Denver and a BA in Finance from Colorado State University.           is full Professor of Business Administration, especially technology and innovation management at the WHU – Otto Beisheim School of Management, in Vallendar, Germany. He is a regular Visiting Professor and a member of the Center for Research in Technology and Innovation at the Kellogg School of Management at Northwestern University, USA. He is also Honorary Professor and Principal Fellow within the Faculty of Business and Economics at the Melbourne Business School, The University of Melbourne, Australia. Professor Ernst has published about 50 papers in the fields of innovation, technology, new product development and intellectual property management in leading journals such as Journal of Marketing, Journal of Product Innovation Management, Research Policy and others. He has won multiple research and best teacher awards. He is also a member of the editorial boards of

xviii

    

Journal of Product Innovation Management, Creativity and Innovation Management and Journal of Knowledge Management. Professor Ernst is co-founder of PatentSight GmbH, a LexisNexis Company, located in Bonn, Germany. He advises and speaks to corporations worldwide and he supports start-ups as a business angel.      .        received her PhD in biochemistry from the University of Wisconsin–Madison. Her professional career spans work in the biotechnology and software industries, with special expertise in network science. She is a cofounder of PatentVector LLC.       .      is a Data Scientist (Addx Corporation) with the Office of the Chief Economist at the United States Patent and Trademark Office (USPTO) and was previously a patent examiner in the field of aerospace systems. Prior to joining the USPTO, Alex served 25 years in the United States Air Force, holding numerous leadership positions in systems development and acquisition, systems engineering and space operations. His research interests include machine learning, innovation policy and technology development. Alex holds an MS in Business Analytics from the George Washington University, an MA in Public Policy from Harvard University, and a BS in Astronautical Engineering from the United States Air Force Academy.          is Development Operations Consultant at the InterAmerican Development Bank, Washington, DC.        .        is Senior Project Leader at GmbH, a LexisNexis company, in Bonn, Germany, and the Friedrich-Alexander-University Erlangen-Nuremberg Germany. Carsten’s research focuses on patent analytics valuation.

PatentSight Lecturer at in Fürth, and patent

       ó  is Associate Professor of Economics at the Autonomous University of Madrid (UAM) and Researcher at the UAM-Accenture Chair in Economics and Management of Innovation. He works regularly as a policy analyst/consultant for international organizations such as the OECD, the European Commission and the World Bank. He sits on the Executive Committee of the European Forum for Studies of Policies for Research and Innovation (Eu-SPRI Forum).

   

xix

              is a licensed attorney in Panama, specialized in intellectual property. In 2018, he worked as a Young Professional at the Intellectual Property, Government Procurement and Competition Division of the World Trade Organization (WTO). In this capacity he supported the work of the WTO Secretariat with matters relating to The Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS), by planning, coordinating and delivering technical assistance and capacity-building activities, and assisting in preparations for TRIPS Council meetings, among other duties. Gutierrez holds an MA in Intellectual Property, Social Networks and New Technologies from ESADE, Spain, as well as a Law degree and an MA in Procedural Law from the Universidad Catolica Santa Maria La Antigua in Panama.             (PhD, MIT Economics) studies innovation, productivity and competition, including R&D productivity in the pharmaceutical industry, specifically the role of geographic and academic spillovers; the firm-specific and policy determinants of the diffusion of new products; generic competition; and the use of markets for technology. Recent work examines the effect of trade and IP policies on the level, location and direction of R&D investment and competition. She currently holds the Chair in Markets for Technology and Intellectual Property at MINES ParisTech and is a member of the Conseil National de Productivité in France. She is an associate editor at the International Journal of Industrial Organization and a Research Fellow at the Centre for Economic Policy. She previously held positions at Carnegie Mellon University, Duke University, London Business School and the Toulouse School of Economics. She has also been a Visiting Scholar at the Federal Reserve Bank of San Francisco, the University of Hong Kong and Northwestern University.             is the IGT Professor of Intellectual Property Law at the William S. Boyd School of Law, University of Nevada, Las Vegas. She previously taught at Florida State University, in both the College of Law and the School of Motion Pictures, Television, and Recording Arts. She received her JD with High Honors from the Duke University School of Law, where she served as Executive Editor of the Duke Law Journal. She simultaneously earned her MA in Philosophy from the Duke University School of Graduate Studies. Prior to teaching, she clerked for Judge Harry T. Edwards of the United States Court of Appeals for the DC

xx

    

Circuit, and practised law for three years with the Washington, DC office of Fried, Frank, Harris, Shriver & Jacobson. In addition to her law review articles and book chapters, Professor LaFrance has authored or coauthored six books, including Intellectual Property Cases and Materials (Thomson West 2018), Understanding Trademark Law (Carolina Academic Press 2019), Understanding Intellectual Property Law (Carolina Academic Press 2019), Global Issues in Copyright Law (West 2009), Entertainment Law on a Global Stage (West 2015), and Copyright Law in a Nutshell (West 2017).        .        , a Professor at the University of California, Davis in the Agricultural and Resource Economics Department, has published research on topics in applied microeconomics ranging from poverty dynamics, climate change and childhood nutrition to technology adoption, intellectual property and innovation policy. Collaborating with researchers, students, NGOs, governments, and private firms, he has lived and worked in India, Haiti, and throughout Sub-Saharan Africa, North Africa and Europe. He currently serves as Co-Editor of the American Journal of Agricultural Economics. He was a Fulbright Scholar in Morocco before earning his MS (2000) and PhD (2004) in Applied Economics from Cornell University.               is an Economic Affairs Officer in the Trade in Services and Investment Division of the World Trade Organization. His sectoral responsibilities include health-related services and many business services. He is also responsible for the Services IntegratedTrade Intelligence Portal (I-TIP services) database. Before joining the Trade in Services and Investment Division in 2014, he was responsible for trade in services statistics-related issues in the WTO’s Statistics Group. He notably contributed to the UN Manual on Statistics of International Trade in Services. He has participated in many technical assistance and capacity-building activities on GATS as well as on trade in services statistics in Africa, Asia, the Middle East, Latin America and South Eastern Europe. Before joining the WTO in 2003 he worked in the OECD for seven years, mainly on services-related issues.      .       is Professor of Economics and former Associate Dean for Social Sciences at the University of Colorado, Boulder, USA. He has been a Lead Economist in the Development Research Group at the

   

xxi

World Bank. He is also a Research Fellow at the Peterson Institute for International Economics, a Fellow at the Kiel Institute for World Economics, and an Adjunct Professor at the University of Adelaide. He has been a Visiting Professor at the University of Bocconi and a visiting scholar at the CES-Ifo Institute at the University of Munich and the China Center for Economic Research at Peking University. He serves also as a consultant for the World Bank and the World Intellectual Property Organization and recently chaired a panel of the National Research Council on intellectual property management in standards-setting organizations.              is Chief of the International Trade Statistics Section at the WTO. He studied economics with a specialization in public finance and statistics/econometrics at the University of Hohenheim and holds a doctorate (Dr.oec.) in economics. In 1990, he joined the United Nations Economic Commission for Europe. He was editor-in-charge of the Statistical Journal of the UN ECE. Since 1994, he has worked with GATT/WTO. His current interests include measuring trade in services flows, trade in value added and digital trade.      .      -     is Counsellor in the Intellectual Property, Government Procurement and Competition Division of the World Trade Organization (WTO), which is responsible for the administration of the Agreement on Trade-related Aspects of Intellectual Property Rights (TRIPS Agreement). In this capacity he gives legal advice in relation to intellectual property and the TRIPS Agreement to members and observers of the WTO and deals with TRIPS-related dispute settlement cases among WTO members. He serves as the Secretary to the TRIPS Council, the governing body of the TRIPS Agreement, and was Secretary to the Special Session of the TRIPS Council, the negotiating group dealing with TRIPS negotiations under the Doha Round. Mr. Meier-Ewert is a graduate of the University of Oxford, UK, where he studied Politics, Philosophy and Economics (PPE) and holds a Law degree from the Ludwig-Maximilian University in Munich, Germany. He has widely lectured on TRIPS and WTO-related matters.             was appointed Deputy Head of CTPA’s Tax Policy and Statistics Division in May 2014. He has over 30 years’ experience as

xxii

    

an economist in both government and the private sector. He has spent the last 20 years at EY where he founded and lead a group of 34 tax policy economists, statisticians and survey specialists. Mr Neubig is EY’s Director of Quantitative Economics and Statistics, where he serves as an advisor to numerous public and private clients on federal, state and global tax policy issues, including revenue and economic impact. He holds a BA degree in Economics from Kalamazoo College and a PhD in Economics from the University of Michigan.           is Principal of the Sir Arthur Lewis Community College in St. Lucia, Senior Advisor on Structural Policies and Innovation at the OECD Development Centre, member of the executive bureau of the UN Committee for Development Policy and former WTO Chair and Senior Fellow at the University of the West Indies.       .   was, from 2004 to 2019, an International Trade Analyst in the Services Division, Office of Industries, at the United States International Trade Commission, a US government agency tasked with providing information and analysis on issues of trade and competitiveness to the US Trade Representative and the US Congress. While at the USITC, Mr Oh wrote on a wide variety of services trade issues but his particular focus was audiovisual services, including online content. He was a major contributor to many USITC reports, including US–MexicoCanada Trade Agreement: Likely Impact on the US Economy and on Specific Industry Sectors, Inv. TPA-105-003 (April 2019); Recent Trends in US Services Trade 2018 Annual Report (June 2018); Global Digital Trade 1: Market Opportunities and Key Foreign Trade Restrictions, Inv. 332-561 (July 2017); Trans-Pacific Partnership Agreement: Likely Impact on the US Economy and on Specific Industry Sectors, Inv. TPA-105-001 (May 2016); Recent Trends in US Services Trade: 2015 Annual Report, Inv. 332-345 (May 2015); Digital Trade in the US and Global Economies, Part 1, Inv. 332-531 (July 2013), and Digital Trade in the US and Global Economies, Part II, Inv. 332-540 (August 2014). While at the USITC, Mr Oh also published working papers and trade briefings on a range of topics, including Nigeria’s film industry, and the growing services sectors in India and Vietnam.         .         is an Economist with the Office of the Chief Economist at the United States Patent and Trademark Office. His

   

xxiii

research interests include the economics of innovation, intellectual property and applied microeconomics. Nick holds a BBA in Economics and Mathematics from the University of Wisconsin Eau Claire, an MA in Economics from Indiana University, Bloomington, and a PhD in Economics from Indiana University, Bloomington.                is Chief of Trade Costs Analysis at the WTO. Her research focuses on trade and trade policy analysis. Her papers have been published in The American Economic Journal and the Journal of International Economics among others and she has authored various books. She is one of the lead authors and coordinators of the World Trade Report (WTR) and WTO-WBG publications on trade and poverty. She is author of ‘A Practical Guide to Trade Policy Modelling’ (WTO 2012) and ‘An Advanced Guide to Trade Policy Modelling’ (WTO 2016). She has served the WTO Dispute Settlement in several Panel and Arbitration cases. Prior to joining the WTO in 2000, she was Lecturer in Economics and Statistics at the University of Southampton and Research Fellow in the research division of Confindustria. She has also taught at the University of Geneva and LUISS. She holds a PhD in Economics from the University of Southampton.               is Professor of Economics at the University of Pavia and also holds a position as Adjunct Professor at the University of Aalborg. Her research is focused on innovation in developing countries, clusters, global value chains (GVC), and foreign direct investment (FDI) and multinationals. She has published widely in international journals and her last book was Upgrading to Compete: Global Value Chains, Clusters, and SMEs in Latin America (Harvard University Press 2007).                    is Assistant Professor Tenure Track in Science & Technology Policy at EPFL (Lausanne, Switzerland). He joined the Institute of Technology and Public Policy at the College of Management of Technology at EPFL in late 2014. Prior to that, he was a research fellow then a senior research fellow at the University of Melbourne (Australia) from 2010 to 2014. He was affiliated with the Melbourne Institute of Applied Economic and Social Research at the Faculty of Business and Economics. Gaétan obtained a PhD in Economics from the Université libre de Bruxelles (Belgium), Solvay

xxiv

    

Brussels School of Economics and Management in 2010. The overarching objective of Gaétan’s research is to provide the policy environment that best addresses the needs of the knowledge economy. This objective is met by providing sound empirical evidence on research questions related, for example, to intellectual property issues, to the measurement of intangible capital, and to higher education systems to name but a few topics of interest. His work has appeared in international peer-reviewed scientific journals, such as Strategic Entrepreneurship Journal, Journal of Economics & Management Strategy, Research Policy and European Economic Review.             is Managing Director at PatentSight GmbH, a LexisNexis company, in Bonn, Germany. He has been involved in patent analytics for more than 10 years. Marco consults for and collaborates with various firms, for example members of the DAX 30 (Germany), Nikkei 225 (Japan) and S&P 500 (USA).          is Counsellor in the Trade in Services and Investment Division of World Trade Organization. With the WTO since 2002, he has been involved in various functions relating to negotiations, technical assistance, dispute settlement, and policy research. He is the Secretary to the Special Session of the Council for Trade in Services. Between 2014 and 2016, he was senior advisor at the Office of the Chief Trade Advisor for Pacific Island Countries, in Vanuatu. Dr Roy has published widely on such topics as trade in services, foreign investment, and regional economic integration. He co-edited the Research Handbook on Trade in Services, published by Edward Elgar (2016), and Opening Markets for Trade in Services: Countries and Sectors in Bilateral and WTO Negotiations, published by Cambridge University Press (2008).         ˊ   ˊ is a Research Economist at the Economic Research and Statistics Division of the World Trade Organization (WTO). She joined the WTO in 2016 and contributed to several publications, including the annual World Trade Reports. She also co-authored an ILO–WTO publication on Investing in Skills for Inclusive Trade and co-coordinated the World Trade Report 2019 on the Future of Services Trade. Stela’s research topics are at the junction of international economic integration, global value chains, and economic growth. She holds a PhD in International Economics from the Graduate Institute of International and Development Studies, Geneva.

   

xxv

              is Senior Researcher in the Innovation Group at the KOF Swiss Economic Institute at ETH Zurich. In his work on empirical innovation economics, he combines firm and regional-level data with patent data. He is an experienced user of patent databases and innovation survey data. His research interests include knowledge spillovers, technology clusters and the internationalization of R&D. He is a project leader of international projects on patent data, including external collaborations with academic institutions in Switzerland and Germany.               is a creative industries consultant and graduate of the Sir Arthur Lewis Institute of Social and Economic Studies, University of the West Indies.          is a PhD graduate of the University of Agder, Norway and CEO of the Copyright Society of Composers, Authors & Publishers Inc. in Barbados.         .      is a Professor of Information Technology and Marketing at Carnegie Mellon University. He received his BSc in Electrical Engineering (summa cum laude) and his MSc Telecommunications Science from the University of Maryland, and received his PhD in Management Science and Information Technology from the Sloan School of Management at MIT. Professor Smith’s research uses economic and statistical techniques to analyse firm and consumer behaviour in online markets – specifically markets for digital information and digital media products. Prior to receiving his PhD, Professor Smith worked extensively in the telecommunications and information systems industries, first with GTE in their laboratories, telecommunications and satellite business units, and subsequently with Booz Allen and Hamilton as a member of their telecommunications client service team. While with GTE, Professor Smith was awarded a patent for research applying fuzzy logic and artificial intelligence techniques to the design and operation of telecommunications networks.             is a partner at the Brazilian law firm Fialho Salles Advogados, focusing on international trade regulations, intellectual property and competition law. He is a PhD candidate in International Law (USP), currently taking part in the WTO PhD Support Program. He

xxvi

    

has an LLM (with merit) in International Business Law (London School of Economics); he also has Executive MBA (Fundação Dom Cabral); Bachelor of Laws (UFMG); and Bachelor of International Relations (PUC Minas) degrees.                 is a Senior Associate at the Brazilian law firm Fialho Salles Advogados, focusing on intellectual property law, data protection and commercial agreements. She has an LLM in Commercial Law (University of Cambridge); she also has Specialization in Intellectual Property Law (University of South Africa/WIPO) and Bachelor of Laws (UFMG) degrees.           is Director of the WTO’s Intellectual Property, Government Procurement and Competition Division. He formerly directed the Global Intellectual Property Issues Division of WIPO (including the Traditional Knowledge Division and Life Sciences Programme), covering IP and genetic resources, traditional knowledge and folklore, the life sciences, and related global issues including public health and climate, the environment, climate change, human rights, food security, bioethics and indigenous issues. He earlier held appointments in the Australian diplomatic service and worked in private practice as a patent attorney, and he has held a number of academic and teaching positions. He has published widely on international IP law and policy, and cognate policy and legal questions. His education encompasses law, international relations, computer science, mathematics, philosophy, classical Greek and theology.            is Professor of Information Systems and Management at the Heinz College at Carnegie Mellon University and at the Tepper School of Business, with an interest in how information and communication technologies, and associated digitization of information impact consumers, business and policies. His research has a particular focus on how digitization (and associated piracy) in copyrighted industries is affecting the incentives of content providers, distributors and users, and the economics of information security and privacy. Dr Telang has published extensively in many top management and policy journals like Management Science, Marketing Science, ISR, MIS Quarterly, Journal of Industrial Economics, Journal of Policy and Management, and NBER chapters. He held senior editor positions at ISR (Information Systems

   

xxvii

Research) and MIS Quarterly. Dr Telang provides extensive consulting and speaking at various firms.       .      is the Chief Economist at the US Patent and Trademark Office (USPTO) and a Research Associate at the Centre for European Economic Research (ZEW). Dr Toole joined the USPTO with experience in the private sector, academia and government. While completing his PhD in Economics at Michigan State University, Andrew Toole was a Senior Economist for Laurits R. Christensen Associates, where he conducted studies on total factor productivity, cost and price analysis and competitive strategy. In 1998, Dr Toole went to Stanford University as a postdoctoral student before becoming a faculty member at Illinois State University and Rutgers University in New Jersey. As an academic researcher, Dr Toole was asked to advise on science and technology policy issues for institutions such as the US National Academies of Science, US National Institutes of Health, and the US Department of Agriculture (USDA). In 2010, he joined the Science Policy Branch of USDA’s Economic Research Service. His research focuses on the economics of innovation, intellectual property, and related science and technology policies. Dr Toole has published in the Journal of Law and Economics, the Review of Economics and Statistics, Research Policy, Management Science, and many other peer-reviewed journals.       .        is Earl B. Shurtz Research Professor at the University of Kansas School of Law, and the winner of a 2015 University Scholarly Achievement Award at the University of Kansas. He received his PhD in Biology from Harvard University in 1997, JD from Harvard Law School in 2000, and BSc from Queen’s University (Canada) in 1991. He joined the University of Kansas School of Law in 2005. Torrance teaches and conducts research in patent law, intellectual property, innovation, food and drug regulation, biotechnology law, biodiversity law, biolaw, and empirical, experimental, and big data approaches to the law. Specific research foci include open, user and collaborative innovation, design, and legal issues surrounding genes, biotechnology, genetically-modified organisms, synthetic biology, conservation biology, and de-extinction.          is a senior staff member of the WTO responsible for telecommunications, ICT services, and electronic commerce. In these areas, she follows policy, regulatory practices, and trends in technology

xxviii

    

and business models. She is responsible for keeping up to date on how ICT trade and e‑commerce developments relate to the WTO and General Agreement on Trade in Services (GATS). She frequently conducts multistakeholder national and regional workshops on e-commerce work taking place within the WTO. For many years, she has also provided technical assistance to telecommunications/ICT regulators on the implementation of GATS provisions (including the disciplines of the GATS Annex on Telecommunications and the Reference Paper commitments on sector regulation). She also helps governments who are negotiating WTO accession or free trade agreements with the technical aspects of drafting their commitments in the sectors she covers. She has worked at the WTO since 1990 and graduated from Columbia University in 1980 with an MA in International Affairs.              was Counsellor in the Intellectual Property, Government Procurement and Competition Division of the WTO from 2001 to 2019, contributing to work on TRIPS and public health, TRIPS-CBD, Patents, Undisclosed information, Economics of TRIPS, IP and Transfer of Technology, IP and Climate Change, and IP and Competition Policy. She currently holds a part-time Adjunct Professor position at the Georgetown University Law Centre, the position of Honorary Professor at the National Law University, Delhi, and has been a member of the Governance Board of the Medicines Patent Pool, a non-profit organization based in Geneva since 2015. Ms Watal holds post-graduate degrees in both law and economics and, prior to joining the WTO, she had more than 22 years of experience in government in India. She represented India at a crucial stage in the Uruguay Round TRIPS negotiations from 1989 to 1990. She is the co-editor of two WTO books: A Handbook on the TRIPS Agreement (Cambridge University Press 2012) and The Making of the TRIPS Agreement (WTO 2015). She has also authored Intellectual Property Rights in the WTO and Developing Countries (Oxford University Press, India and Kluwer Law International, 2001) and several peer-reviewed journal articles on issues related to the law/economics of intellectual property rights.             is Professor of Contract Law and of Intellectual Property at the Law School of the University of

   

xxix

Geneva, Switzerland. He authored a doctoral thesis in Swiss and comparative copyright law, which he completed as a Visiting Scholar at the Max-Planck Institute for Intellectual Property, Competition and Tax Law in Munich in 1996. He then practised law in Switzerland, before obtaining an LLM degree from Columbia Law School in New York City in 2001 and being admitted to the New York bar in 2002. Jacques researches, publishes and discusses on topics related to various aspects of intellectual property law, contract law, particularly on the commercialization of intellectual property assets with the use of transfer of technology, licensing and franchising, IT and Internet and digital law, as well as alternative dispute resolution mechanisms for IP and technology disputes.            -        is Head of the Composite Indicator Research Section, Economics and Statistics Division, and Co-Editor of the Global Innovation Index (GII) at the World Intellectual Property Organization (WIPO). He joined WIPO as Senior Economist in 2010 to help set up WIPO’s economics work under the Chief Economist, including the World Intellectual Property Report and the GII. Before joining WIPO, he was an Economist and Co-Leader of the Innovation Strategy Project at the OECD Directorate for Science, Technology and Industry. Prior to that, he was the Swiss National Science Fellow at the Berkeley Center for Law and Technology, University of California, Berkeley, and the Peterson Institute for International Economics, Washington, DC. He is currently preparing a book on ‘Harnessing Public Research for Innovation in the 21st Century: An International Assessment of Knowledge Transfer Policies’ with Anthony Arundel and Suma Athreye for Cambridge University Press. He holds an MA in International Economics from the Maastricht Economic Research Institute on Innovation and Technology, University of Maastricht, and a PhD in Economics from the University of St Gallen, Switzerland. He teaches international economics at Sciences Po Paris and the World Bank Institute.     .   is Regents Professor of Law and Communication and Director of the Center for Law and Intellectual Property at Texas A&M University. Born and raised in Hong Kong, he previously held the Kern Family Chair in Intellectual Property Law at Drake University Law School and was Wenlan Scholar Chair Professor at Zhongnan

xxx

    

University of Economics and Law in Wuhan, China. He served as a Visiting Professor of law at Bocconi University, Hanken School of Economics, Hokkaido University, the University of Haifa, the University of Helsinki, the University of Hong Kong and the University of Strasbourg.       .   is a Senior Economist with the Center for Economic Studies at the US Census Bureau. Nikolas’s research interests are in innovation, intellectual property and technology transfer. His work has been published in Research Policy, Science, Journal of Regional Science and more. Nikolas received his PhD from the University of California, Davis. Prior to receiving his PhD, Nikolas started a non-profit agency and worked in finance.

PREFACE

It is now more than 25 years since the WTO TRIPS Agreement entered into force, setting intellectual property (IP) standards at the centre of multilateral trade rules. The quarter-century since then has seen a fundamental transformation in the scale, diversity and very nature of crossborder commercial transactions in knowledge and knowledge products. An array of technological, economic, social and policy factors has driven this transformation and diversification. The disruptive impact of technological change – the prospects that it brings for sustainable development and for a more equitable world, along with concerns about its potential to displace and disenfranchise, and to entrench inequities – has placed it at the centre of policy debates and practical initiatives about economic and social development, and indeed a host of wider public policy issues today. The framing of the United Nations Sustainable Development Goals (SDGs) in 2015 – a blueprint for international cooperation and national action towards an ambitious set of targets in 2030 – was striking for the high degree of recognition of the need for the development and implementation of new technologies to address fundamental development needs. Innovation, as such, was recognized as a development goal in itself. As international officials charged with providing policy support to developing country members of the WTO, we have been increasingly struck by the demand from a wide range of policymakers and officials for a fresh, up to date and inclusive information platform to support a truly contemporary discussion and practically relevant planning of technical cooperation that reflects the current understanding of the economic, legal and policy aspects of trade in knowledge today. The TRIPS Agreement was concluded in the absence of any understanding of the dramatic impact of the development of the Internet and increasing global access to digital networks. Indeed, it is a piquant historical irony that Tim Berners‑Lee was in the process of inventing the World Wide Web at CERN, in Geneva, a few minutes away from the negotiating rooms where the text of the TRIPS Agreement was being hammered out at exactly the xxxi

xxxii

   

same time. Yet TRIPS did not account for the revolutionary upheaval in the way knowledge would be created, disseminated and traded once the Web had reached the wider community; its negotiating roots could be found in attempts in the 1970s and 1980s to develop a code on counterfeit trade, with a concentration on IP as part of the added value in traded physical goods, rather than as a tradable good in itself. As an agreement on ‘trade-related aspects’ of intellectual property rights, TRIPS set the stage for the impending fundamental transformation of the relationship between IP and trade. The volume is intended to serve as a conceptual and empirical foundation for a renewed set of policy discussions, capacity building and technical assistance for governments seeking to engage with the knowledge economy for development, through new trade, business and employment opportunities. It seeks to review the legal character and economic implications of international transactions that facilitate the transfer or diffusion of new knowledge and intangible content, through traditional trade channels and new forms of business and knowledge transfer. To provide a complete picture, we have aimed to include legal, policy and economic analysis in a coherent manner. Thus we sought accounts of how knowledge crosses international borders in different ways, and how to measure these flows, as well as the analysis of legal and policy scholars reviewing the evolving laws, regulations and policies that govern such trade, and analysing the legal character of knowledge transactions in today’s international economy. In a call for papers, we set out several broad research questions: 1.

2. 3.

How can we measure different forms of global cross-border flows of knowledge and knowledge products, including digital content? What can we say from the relevant data about the scale, and geographical and time trends (for example, pre-TRIPS and post-TRIPS) of such flows? What legal and policy questions are raised by current trends in cross-border knowledge flows? What can we say about the effects of knowledge flows on economic development, growth and productivity; domestic innovation; technology transfer; trade, especially high-tech trade and trade in digital content; and the development opportunities afforded by new avenues for trade and employment?

     

4.

5.

6.

xxxiii

What is the relationship of such knowledge flows to improved creation of and access to global public goods, especially technologies for agriculture, health and climate change, and (information and communication technology) ICT? What can we say about the legal and economic character of trade in digital content covered by intellectual property rights (IPRs), including issues arising from global or regional regulatory frameworks, new business models for content distribution, and cross-border licensing and enforcement? What is the current experience in specific sectors, and in markets for IPRs as such, and what insights does this experience offer policymakers?

We were delighted at the strong, positive response elicited by this call for essays, and the willingness of the leading international scholars whom we invited to contribute. While we cannot expect this broad and ambitious set of questions to have been definitively answered, we believe that the scholarly work collected in this volume should establish a stronger and more up-to-date empirical, theoretical and methodological basis for a vital continuing international conversation centred on these questions. We hope that policymakers and scholars will continue to return to these questions, armed with greater insights and greater curiosity sparked by an encounter with this book. The book project was greatly enhanced by several peer-review sessions, in particular the Seminar on Intellectual Property and Knowledge Flows in a Digital Era, convened at the WTO in November 2018, as a follow up to the call for papers. This Seminar drew together over 40 policymakers from developing and least developed countries around the world, who were joined by 20 internationally renowned economist lawyers and policymakers. Their discussions on how to map and assess the impact of knowledge flows across borders and charting their legal, economic and policy dimensions helped inform and consolidate the scheme for this book and served to strengthen and make more practical and relevant the contributions that are now gathered together in this collection. The Seminar considered five broad themes that in turn helped to shape this book: mapping the interface between trade, intellectual property rules and knowledge flows; measuring trade in knowledge; the impact of knowledge flows on trade and development; policy, regulatory and legislative frameworks; and the way forward on trade rules and economic implications for cross-border knowledge flows.

xxxiv

   

Addressing the Seminar, the WTO Deputy Director General responsible for this field, Mr Xiaozhun Yi, remarked that The TRIPS Agreement was a landmark recognition of the importance of the knowledge component of trade. Yet in the years since conclusion, digital disruption has utterly transformed the interplay between trade and knowledge. This radical transformation creates a compelling need to update our understanding of the context for TRIPS rules and the intellectual property system within the framework of trade and development policy. We need to fundamentally update both our theoretical understanding, and the empirical, factual base we work from. This is important, too, because in this period, development policy has laid increasing emphasis on the knowledge component of trade, and has raised practical questions of how developing economies can make best use of the opportunities provided by the knowledge economy: SDG 9, notably, identifies innovation as such as a goal for sustainable development. The WTO’s technical assistance activities have sought to respond to evolving demand from developing country Members for capacity building in these areas, but in a relatively ad hoc way, largely as an adjunct to existing programme structures, and these activities have demonstrated the unmet demand and practical need for more systematic capacity building on an updated base of information.1

This demand follows the disruptive effect of technological change which has impacted significantly on the area of trade involving IP rights, and the exercise of policy options under the TRIPS Agreement, opening up new avenues for development as IP in itself becomes a tradable good and IP plays a pivotal role in dispersed international production chains and in facilitating knowledge transfer. Policymakers therefore confront an entirely new set of challenges integrating TRIPS measures into trade policies that respond to a digitally transformed knowledge economy; these challenges extend to basic capacity to measure and to map the IP dimension of trade, and thus to develop an integrated understanding of how the IP system, and new forms of trading in knowledge, can function in the contemporary international economy to service diverse national development priorities. Recognizing the need for this initiative to retool and rebase the WTO’s technical assistance in the TRIPS area, responding to the fundamentally transformed technological and trade landscape, the themes of this book were also the subject of a focused session at the WTO’s Aid for Trade Global Review 2019. Hence the initiative for this book was elaborated in 1

See www.wto.org/english/news_e/news18_e/trip_08nov18_e.htm.

     

xxxv

dialogue with policymakers and officials from across the developing world and is intended to provide a scholarly, but practical and forward-looking, resource as the foundation for a new generation of technical assistance and policy dialogue. This collaborative and consultative background, but more importantly the remarkable quality, timeliness and cross-disciplinary character of the authors’ contributions, inspires hope that this volume – and the associated resources at www.wto.org/trade-in-knowledge – will indeed help to catalyse a fresh, contemporary approach to policy development and technical assistance in the complex, but critical, area of today’s trade in knowledge, where the IP system and knowledge flows interact in ever more diverse ways. We are profoundly grateful to the authors who have so generously devoted their time and efforts to produce invaluable new scholarship and policy insights. Our opening chapter seeks to outline each of their contributions and to set them into the broader framework of this book and the aspirations we have for it, and we trust that this will encourage the reader to explore their work in full, beyond the chapters contained here, as among the roll of authors are many who have made a major contribution to scholarship and policy development. We count ourselves as most fortunate to have among these authors invaluable colleagues who are also new or long-established friends, and we thank them for their support and their patience as this project moved forwards. For his encouragement and support, we thank Deputy Director‑General Yi, and also record our warm thanks to our colleagues, Anthony Martin and Heather Sapey-Pertin of the WTO’s Information and External Relations Division for their advice, support and patience, Karyn Russell who provided cheerful and efficient administrative support throughout the editorial process and Giovanni Bertinelli and Francesco Hernandez Fernandez for their scrupulous attention to final editorial corrections. At Cambridge University Press, we are indebted to Kim Hughes for her encouragement and support at a critical early stage. Co-editors Antony Taubman Jayashree Watal

1 Thematic Overview: Charting the Evolution of Knowledge Flows                            

Introduction: Trade in Knowledge Today The Nexus: Knowledge Flows, Trade and Intellectual Property Understanding cross-border flows of knowledge, often associated with transactions involving intellectual property (IP), is essential to analysing how modern economies grow and evolve, and how international trade can underpin technological development. How to stimulate knowledge flows and make more effective, systematic use of them is an immediate practical concern for contemporary policymakers and analysts seeking to frame and implement policies for economic and technological development, that strengthen innovation systems and tap into indigenous creative and innovative capacity. Trade is understood to serve as a major conduit for the knowledge dissemination and technology spillovers that are essential for sustainable development today. And the IP system has been crafted and implemented ostensibly to facilitate both innovation and the dissemination of the fruits of innovation. Trade in knowledge as such – transactions specifically over the licensing or transfer of IP rights, for instance – has become a practical reality and a major source of dynamism and disruption. The complex and dynamic interaction between the IP system and international trade is therefore critical to our understanding of knowledge flows and their contribution to development. Yet technological change – digital disruption – is today itself increasingly a factor reshaping and diversifying the ways in which knowledge is developed, managed, transacted and disseminated. Accordingly, in the quarter-century since the WTO was established, and since its Agreement on Trade-Related Aspects of Intellectual Property Rights (‘TRIPS’) came into force, both the knowledge dimension of trade and the functioning of the IP system have been radically transformed. Many salient aspects of 



    

these developments have been closely studied by scholars. Yet, when surveying available material to respond to WTO members’ increasing interest in understanding these complex phenomena, we found few resources that look at knowledge flows, trade patterns and the evolution of the IP system in an holistic way, while responding also to the development implications of technological change. Hence the WTO Secretariat initiated a dialogue between policymakers and analysts, including a call for papers, several workshops and a peer review process, to fill this gap in available materials, an effort culminating in the present edited collection of readings and cutting-edge analysis. This volume forms a central part of the Secretariat’s efforts to build an up-to-date, inclusive and empirically well-founded information platform to support policy development by its members in this critical but challenging area.

Structure The book is organized in four substantive parts, corresponding to four dimensions of the need for systematic understanding of cross-border knowledge flows or ‘trade in knowledge’: (i) an overview of the conceptual framework for trade, IP and international knowledge flows (ii) possibilities and challenges for measuring trade in knowledge (iii) the impact of knowledge flows on trade and development (iv) considerations for the governance frameworks that apply to these knowledge flows. The present chapter provides a general introduction to the book, along with a thematic overview of the individual chapters contained in each of these parts. A concluding chapter then aims to draw together observations and insights from the substantive chapters and to offer ideas for future directions in research and policy dialogue.

Background Knowledge flows, and their trade and development impacts, have come to the forefront of contemporary trade policy and trade relations. Governments seek to clarify and redefine their economic and development interests in the light of the disruptive effects of technological

  



change, the changing patterns of production and trade, and fundamental shifts in innovative and technological capacity. The qualitatively more diverse and quantitatively higher cross-border knowledge flows since the 1990s have in part been associated with a process of diversification of production capacity, associated with the rapid development of emerging economies, notably in Asia. This evolution was, in turn, partly driven by the ICT revolution which so dramatically improved rapid and secure communications that it enabled the rise of global production chains or value chains – a transformation that Baldwin has termed ‘the second unbundling’, as physical steps in the production process (following the ‘first unbundling’, which had separated consumers from producers). As he has phrased it: Globalization accelerated again from around 1990, when the information and communication technology (ICT) revolution radically lowered the cost of moving ideas. This launched globalization’s next phase—call it the ‘second unbundling’ since it involves the international separation of factories. Specifically, radically better communications made it possible to coordinate complex activities at distance. Once this sort of offshoring was feasible, the North–South wage gap that had arisen during the first unbundling made it profitable. The offshoring of production stages to low-wage nations changed globalization, but not just because it shifted jobs overseas. To ensure that the offshored stages meshed seamlessly with those left onshore, richnation firms sent their marketing, managerial, and technical know-how along with the production stages that had been moved offshore. As a consequence, the second unbundling—sometimes called the ‘global value chain revolution’—redrew the international boundaries of knowledge. The contours of industrial competitiveness are now increasingly defined by the outlines of international production networks rather than the boundaries of nations.1

Perhaps not entirely coincidentally, economists – notably Paul Romer – were at the same time recognizing that knowledge was a key and endogenous parameter in a country’s economic development, and could no longer be considered an external factor when modelling economic growth. As he observed at a 1992 World Bank conference on development: All too often, economists concerned with the economy as a whole have been willing to treat the economics of ideas as a footnote to the rest of

1

Richard Baldwin, The Great Convergence, Cambridge, MA: Harvard University Press, 2016, 5–6.



     economic analysis – important for understanding some of the details but not something that changes how we think about big policy questions. A neoclassical model with perfect competition and exogenous technological change continues to frame many, if not most, policy discussions of growth and development. Ideas are routinely ignored . . . ideas are extremely important economic goods, far more important than the objects emphasized in most economic models. In a world with physical limits, it is discoveries of big ideas (for example, how to make hightemperature superconductors), together with the discovery of millions of little ideas (better ways to sew a shirt), that make persistent economic growth possible. Ideas are the instructions that let us combine limited physical resources in arrangements that are ever more valuable.2

He argued at that time that it is necessary to ‘take seriously the economic opportunities presented by the potential for producing new ideas and for diffusing existing ideas to the widest possible extent. In so doing, we must recognize that ideas are economic goods which are unlike conventional private goods and that markets are inherently less successful at producing and transmitting ideas than they are with private goods’.3 Just at the time that such economic analysis was emphasising the necessity – and the significance for policymakers – of incorporating knowledge into models of growth and development, trade negotiators, led by the then industrialized countries, negotiated the TRIPS Agreement in order to increase legal certainty and predictability in the transfer of knowledge. Formally concluded in 1994 and entering into force the following year, its text had been essentially settled by 1992.4 The Agreement laid out minimum standards for the protection, use and enforcement of intellectual property rights that all WTO members would have to follow, while also incorporating provisions on technology transfer for least-developed countries and express scope for policymakers to craft laws to facilitate the flow of knowledge. The incorporation of the TRIPS Agreement into the package of international trade law that came into effect upon the formation of the WTO in 1995 meant that it 2

3 4

Romer, Paul M., ‘Two Strategies for Economic Development: Using Ideas and Producing Ideas’, Proceedings of the Annual World Bank Conference on Development 1992, Supplement, Washington, DC, World Bank Economic Review, 1993, at 63. Ibid., 89. For individual negotiators’ accounts of the TRIPS negotiations, and a thematic overview of the negotiations, see Jayashree Watal and Antony Taubman, The Making of the TRIPS Agreement: Personal Insights from the Uruguay Round Negotiations, Geneva: WTO, 2015, available at: www.wto.org/english/res_e/booksp_e/trips_agree_e/history_of_trips_nego_e .pdf.

  



comprised a core element of the emerging legal and economic framework for what were to be the largest global transfers of knowledge in the period from the 1990s to the present day. The period since 1995 has seen a fundamental transformation not only in the character and impact of digital technologies, but also in the geographical distribution of innovators and associated IP right owners, especially patents. Over this period, due specifically to the digital revolution, creative cultural content that used to be traded as physical goods are now sold on digital platforms online directly to markets across the globe – a burgeoning ‘trade in knowledge’ that frees valuable content from the physical carrier media once relied upon to transport it. Such ‘digital disruption’ offers new opportunities for development, as small traders and creative industries in the developing world can potentially overcome traditional obstacles to trade, although significant challenges remain, with concerns about the potential effects on beneficial competition of dominant online platforms and the difficulty of establishing alternative avenues for knowledge trade. Equally, the management of IP and IP licensing transactions constitute significant elements of dispersed global production chains and play a key role also in the knowledge spillovers resulting from IP-based transactions.

Conceptual Framework Part I of this book sets out the conceptual framework about what we consider to be trade in knowledge in relation to trade in goods, services and intellectual property. We note the growing importance of trade in intangibles, distinguishing the delivery of knowledge-intensive services from the charges for the use of IP such as licence fees and royalties. We also note the rise in trade of digital products and services, and the IP dimension of such trade, in view of the fact that much of this trade is constituted as transactions in IP rights, whether through the purchase of licences or the transfer of ownership. Finally, this part looks at the existing IP framework for trade in knowledge, analysing both the TRIPS Agreement itself in its contemporary context, as well as provisions in subsequent trade agreements that have bearing both on IP in the digital environment and on trade in digital products; it then broadly points to the emerging policy issues and possible gaps in the regulatory framework thus far. Antony Taubman’s introductory chapter sets the framework for the book. He charts the evolution and diversification of trade in knowledge



    

that has taken place in the quarter-century since the TRIPS Agreement came into force. Entirely new markets have come into being and, arguably, the very character of ‘trade’ is in need of reconsideration. For instance, the disruptive effect of digital technology has meant that much of the content formerly conceived of as ‘added value’ embedded in physical carrier media, traded and measured as ‘goods’, is now traded in the form of specific licences that use IP rights covering the content that is now increasingly accessed online in digital form. Taubman outlines how these new forms of exchange in valuable intangible content confront fundamental assumptions about the nature of trade and its interaction with the IP system, forcing a rethink of what constitutes the ‘traderelated aspects’ of intellectual property. The issues examined include the principle of territoriality of IP rights and the segmentation of markets according to national jurisdictions; the structuring of cross-border commercial exchanges into the two discrete categories of ‘goods’ and ‘services’; the emerging disparity in regional trade agreements between provisions on digital IP standards and on digital products and e-commerce; and the significance of IP rights being treated as assets in investment treaties. The chapter concludes that – whatever formal or legal overlay is applied to these new trading arrangements – it is essential to understand that this is now trade in IP licences as such, rather than trade in goods that have an IP component as an adjunct or ancillary element. Just as Romer and others demonstrated the need for economic growth theory to incorporate intangible knowledge as an endogenous factor, rather than maintaining it as exogenous to models of growth, trade policy must similarly work to incorporate an understanding of the trade in IP licences itself within cross-border commercial exchanges as an integral element of international trading relations. This means treating the exchange and licensing of IP rights systematically and effectively as ‘endogenous’ to trade. This is essential for an accurate empirical picture of trade relations today, given the economic significance both of dispersed global value chains and of trade in ‘pure’ IP content as such, particularly in the creative sectors. Lee Tuthill, Antonia Carzaniga and Martin Roy look at the ways digital technologies have stimulated the information component of services trade and consequently enhanced trade in both goods and services that embody knowledge. They illustrate the role that ICTs have taken on as conduits for digital and digitally-enabled trade. Their chapter briefly describes six important digital developments, namely the cloud, data analytics, (so-called ‘big data’), Internet of Things (IoT), artificial

  



intelligence, robotics and three-dimensional (3D) printing. These technologies are transforming the tradability of services, increasing the growth of cross-border services supplied electronically, across a wide range of sectors, from medical, to educational, financial, audiovisual or professional services. This chapter covers, in broad strokes, the landscape of policy challenges that governments confront as they seek to adapt, including policies that could potentially disrupt the growth of crossborder digital trade such as localization of data. Finally, the chapter provides illustrations of ways negotiators of trade rules have begun to shape new legal frameworks via regional trade agreements (RTAs) that elaborate upon and, in some respects, extend beyond existing multilateral trade rules. For example, in some cases, WTO members with no GATS commitments in basic telecommunication services made RTA commitments on the sector, with no limitations. However, trade disciplines allow countries to take measures to pursue legitimate policy objectives, provided that they are not applied in a manner that would constitute a means of arbitrary or unjustifiable discrimination or a disguised restriction on trade. Wolf R. Meier-Ewert and Jorge Gutierrez explore how regulatory responses to emerging IP issues in digital trade may develop at the international level. In particular, the authors examine how existing mechanisms might influence the chances of developing internationally agreed rules in this regard. The authors note that the primacy of state sovereignty in intellectual property up to the late nineteenth century gave way to the important World Intellectual Property Organization (WIPO) treaties, which still retained some independence of member states and based international regulatory responses directly on national experience. While more regulatory sovereignty was ceded in TRIPS, the WIPO Copyright Treaty and the WIPO Performances and Phonograms Treaty, the adoption of non-binding instruments (such as the WIPO Joint Recommendations in the area of trademarks) show the limits of decision-making by consensus. International non-state solutions such as the Uniform Domain-Name Dispute-Resolution Policy (UDRP) established by the Internet Corporation for Assigned Names and Numbers (ICANN) have introduced separate, technically determined solutions to specific IP issues. Proliferating free trade agreements (FTAs) have emerged as a new platform to agree to IP-related regulatory responses that can be used to project the national solutions of a few dominant FTA partners. However, these FTAs have also served to give legally binding status to internationally agreed non-binding recommendations.



    

The authors go on to highlight how these diverse approaches are apparent in recent IP-regulatory responses to emerging digital issues that are particularly relevant for digital business models, including inter alia Internet service provider (ISP) liability, ‘safe harbour’ provisions and the issue of orphan works, where there appears to be less agreement. Still further away from reaching any kind of agreement are the emerging issues of online exhaustion, data mining and IP-related questions of artificial intelligence.

Measuring Trade in Knowledge Part II of this book discusses the possibilities and challenges for measuring trade in knowledge and what these measurements tell us about recent trends in cross-border knowledge flows. It benefits from chapters contributed by some of the leading researchers in the field. Their work measures both absolute and relative cross-border flows, and trends both in terms of dominant technologies as well as geographical distribution using different metrics. The two key sources of data explored are the balance‑of‑payments (BoP) statistics collected by the International Monetary Fund (IMF), and patent statistics. Andreas Maurer and Joscelyn Magdeleine focus on an analysis of the statistics on IP-related transactions recorded as goods and services in international balance-of-payments data – in effect, an incomplete measure of trade in IP as such, although a measure with its shortcomings. Their chapter opens this part of the book because this has been the standard means of measuring cross-border knowledge flows through trade:5 accordingly, it provides the baseline for improved – more accurate and comprehensive – measurements of cross-border payments for charges for the use of IP, and the authors helpfully identify how statistics could be made more complete and accurate. They chart how licence fees for the use of IP have increased more than four-fold since 1995, twice as fast as exports of high-tech goods and 1.5 times faster than the overall increase of trade in commercial services. This trade, which is driven in part by global production arrangements, remains highly concentrated geographically, with the US and the EU (28) accounting for more than 76 per cent of the receipts in 2016. However, they note how the data suffer from serious lacunae in that only a very small number of mainly 5

See https://data.worldbank.org/indicator/BM.GSR.ROYL.CD.

  



OECD countries report complete information, and there are several definitional problems that prevent a full understanding of the dimension, direction and other details of trade in IP-related products. The authors conclude by making several recommendations to consider in order to improve the statistical basis, including the creation of a new international dataset covering international IP-related flows. Throwing some doubts over the conclusions that can be drawn from an uncritical view of the BoP statistics described in the previous chapter, Thomas Neubig and Sacha Wunsch-Vincent show how tax planning by multinational enterprises has seriously distorted the measurement of cross-border IP flows and affected national measurements of imports, exports, GDP and productivity. The main data sources assessing these flows are described and the main distortions to these flows are assessed. Focusing on the implications of differing taxation regimes, the chapter describes the tax effects on cross-border IP measures in more detail and measures the magnitude of and trends in the tax distortions. The authors estimate that overall tax-induced mis-measurement could be more than 35 per cent of the Charges for Use of Intellectual Property (CUIP), being more for individual countries, particularly those with high tax rates. The authors suggest some initial possible approaches to reducing mismeasurement distortions. Elucidating general patterns of the more valuable technical knowledge flow across borders from patent citations and patents network data, Andrew W. Torrance, Jevin D. West and Lisa C. Friedman test three specific hypotheses designed to shed light on such flows. The first is general, and is inspired by a growing interest in technology transfer between the developed and developing world: the net flow of more valuable technical knowledge tends to be from more developed to less developed countries. It transpires that the importance of citing developed world patents is much greater on average than citing developing world patents throughout the entire time period from 1985 to 2018. From 2010 onwards, the disparity narrows somewhat, in large part due to China. The second hypothesis concerns the trajectory of pharmaceutical knowledge flow between developed and developing countries. Developing world pharmaceutical patents cited by developed world patents also exhibited a general rising trend. However, the average importance of cited developing world patents remained consistently far above that of the developed world level, again in no small measure due to China. This may mean that while smaller numbers of pharmaceutical patents are filed by developing countries, these tend to be more important inventions.



    

Overall, pharmaceutical patent trends in citations, aggregate importance and mean importance do not appear to diverge meaningfully from those for patents across all technological fields. The third hypothesis concerns the trajectory of climate change mitigation knowledge flow (especially from developed to developing countries) after the 1997 Kyoto Protocol and the 2015 Paris Agreement on climate change. Looking only at windpower patents, the authors reject conclusions from earlier more general studies and tell us that there is no clear evidence that the Kyoto Protocol influenced any of the metrics used to characterize wind technology patent knowledge flows but there are some positive trends after the Paris Agreement, which, given how recent these data are, should be treated with great caution. One striking observation apparent from the analyses in this chapter is that US patents make up a huge proportion of patent citations, patent importance and technical knowledge flow within the developed world. Similarly, China has recently come to dominate patent citations, patent importance and technical knowledge flow within the developing world. Laurie Ciaramella, Gaétan de Rassenfosse and Florian Seliger provide a long-term view on the sources of knowledge flow between developed and developing nations. The authors rely on patent data to explore three potential sources of cross-border knowledge flows: R&D collaboration, technology sourcing and technology transfer. R&D collaboration is measured by looking at patents that are co-invented or co-applied for by parties in two different countries. Technology sourcing is measured by patents that are applied for by a party in one country but invented in another country. Technology or knowledge transfer is measured using information on the transfers of ownership (i.e. sales) of patents. All three sources provide very consistent conclusions. First, knowledge flows with East Asia, particularly China, are occurring more frequently. Second, knowledge flows are increasingly concentrated in information and communication technologies. Third, while the United States and Canada traditionally have greater patenting activity with Asia than with Europe, the share of activity between Europe and Asia has been increasing in recent years. Fourth, greater patenting activity between the United States and Canada and Asia implies that the US/Canada region is more likely to benefit from reverse knowledge flows as China progresses towards becoming a technological leader. Jacob Dubbert, Alexander V. Giczy, Nicholas A. Pairolero and Andrew A. Toole provide a valuable survey of empirical studies on cross-border trade in knowledge that use IPRs data. This review finds that most studies

  



use IPRs data, primarily patent documents, as direct indicators of innovation, technology transfer and knowledge flows. Most of the direct indicators are constructed using patent counts, patent citations, licensing and royalty payments, and patent families. The authors also find IPRs data are also used in indirect ways to characterize the relationship between two or more entities (e.g. countries, industries, firms) and serve to measure the degree of technology transfer or knowledge flow. Studies also use IPRs data to construct indexes of country-level IPRs strength, various metrics of IPRs intensity, technological proximity and technology transfer or knowledge flow through inventor mobility across countries. The authors conclude that IPRs data are a valuable source of information for learning about technology transfer and knowledge flows across borders. These data can help to inform the evidence base for policymaking so long as decision-makers are aware of the limitations of such data, such as territoriality, missing contexts and variability, so that not too much is read into the findings. Erick S. Oh focuses in his chapter on recent developments in the digital content industry. He opens his review with a brief discussion of the data limitations that challenge the analysis of this market, and continues with overviews of the key characteristics, important trends and competitive landscapes in each of the major digital content sectors – video games, video on demand (VoD), digital music, and e-publishing (primarily e-books). He describes how more than half of the global digital content market is for video games, China surpassing the US in 2017 as the largest contributor to this sector. The next most important sector is VoD, where the US has about 60 per cent of the global market, thanks to the domination of a few US companies. While this market is getting increasingly competitive, the new entrants are again American entities. China is the second largest consumer of VoD, with that market mainly serviced by Chinese companies. The last two sectors are smaller and their growth is hampered to some extent by copyright infringement issues. Oh also explains how artificial intelligence has been used to good effect to increase market penetration in these different areas. The chapter ends with a short summary of intellectual property rights issues and barriers that are currently affecting the global digital content industry, providing a valuable conceptual framework overview for the issues that are dealt with in Part III of this volume, such as the chapter by Smith and Telang. Vito Amendolagine, Cristina Chaminade, José Guimón and Roberta Rabellotti measure cross-border knowledge flows through yet another



    

variable, namely foreign direct investment in R&D. The authors provide an overview of R&D FDI during the period 2003–2017, analysing key trends of greenfield investments and acquisitions by region and industry. Their main finding is that R&D FDI is highly concentrated in a few large emerging countries such as China and India, and in just a few regions and cities within these countries. Such FDI is also confined to a handful of industries, notably ICT and chemicals. The authors build their analysis on the assumption that learning opportunities for low- and middleincome countries (LMICs) are larger when multinational enterprises (MNEs) collaborate in innovation with local firms, universities and public research institutes, and focus on inward R&D FDI as a driver of economic development and catching-up for LMICs. Their chapter delves into the motivations and implications of this type of FDI, building on a critical review of the existing literature. Finally, the authors provide suggestions for a future research and policy agenda, focusing on the opportunities and challenges for LMICs. Holger Ernst, Carsten C. Guderian and Marco Richter examine the metric of mergers and acquisitions to provide insights into cross-border knowledge flows. They demonstrate how novel patent analytics based on standardized international patent data available to policymakers and managers can help trace specific knowledge flows via mergers and acquisitions. Mergers shape both current and future knowledge flow, inter alia, by affecting the incentives and capabilities of the merged firm to innovate. The European Commission’s decision to approve the Dow/ DuPont merger shows how to measure relevant knowledge flows by means of novel patent analytics, moving beyond conventional patent indicators. Such patent analytics are used in this chapter to trace and assess cross-border knowledge flows. Their analysis – consistent with other chapters in this part – finds that high-quality patents are still largely developed in the US and Europe. However, Asian countries such as China, Japan and the Republic of Korea are catching up rapidly in terms of patent strength and patent quality. The destinations of global patent flows have also changed significantly over time. The authors also show that China has become an important destination for patents, reflecting its increasing role as an important knowledge destination market for many countries, especially Germany. They demonstrate how policymakers and managers can use future-oriented, quantitative patent analytics to improve policy decisions related to cross-border knowledge flows and innovation.

  



Impact of Knowledge Flows on Trade and Development Part III of this book deals with the impact of knowledge flows on trade and development. Some chapters are focused on trade and more particularly on what we can learn from the effect of knowledge inflows on trade and also the effect of trade on creating more knowledge, particularly highlighting the case of global production chains. Others provide specific policy guidance to LMICs on what they should do to further enhance inflows of knowledge. We begin this Part with a comprehensive survey of the literature from two of the leading scholars on technology flows who tell us what economists have found so far with respect to the theory and empirical evidence on the impact of international knowledge flows on international trade, growth, productivity and innovation of the recipient country. Keith E. Maskus and Lee Branstetter consider the key lessons presented by more than two decades of past and ongoing empirical research on international knowledge flows, including their own earlier work (which itself is a leading contribution to scholarly understanding of the field). They recognize the disparities in the emerging and developing world between those countries with sufficient size, economic dynamism and the necessary complementary skills to learn deeply from the global stock of knowledge, and those countries without such advantages. The insights gained help them explain why the income and productivity gains from information globalization do not diffuse as widely as basic models would suggest. The important question of how R&D investments have been globalized, primarily by MNEs, and the implications for knowledge flows are discussed. They demonstrate the crucial role played by investments in human capital and the acquisition of technological capabilities in technology receiving countries. Empirical economic research amply demonstrates that knowledge acquisition and use increasingly requires complementary technical skills within an economy. Achieving broad skill endowments may not be feasible in poor countries for some time, suggesting that investing in deeper skills that may be deployed in global information networks could be a better development strategy. The authors also draw lessons from interesting experimental work aimed at improving education delivery in India through the use of artificial intelligence (AI). Together, these insights point to the need for a broader understanding of how knowledge is transmitted and could be better absorbed in the current era. Maskus and Branstetter conclude with some



    

policy inferences, including that the measured effect of local IPR (and other policies) on the actual incidence of technology transfer tends to be highest in the richer developing countries because these are precisely the places where local skill levels support a significant degree of technology transfer. Margaret Kyle and Mercedes Delgado focus on cross-border knowledge flows through international trade in the post-TRIPS era. They examine trade patterns following the introduction of intellectual property (IP) rights, with a specific focus on a country’s compliance with TRIPS. The authors compare how imports and exports of IP-intensive products changed versus a control group of products for which IP is less critical. Using data from 1993 to 2016 for 151 countries, they show that both exports and imports of IP-intensive goods increased after TRIPS compliance relative to a control group of products. Reflecting the diversifying global landscape for innovation and technology diffusion, in contrast with broad assumptions of a north–south divide, this increase was not confined to high-income countries: developing countries also benefitted and increased their imports of high-IP products from the most innovative countries. They conclude that if trade is associated with knowledge diffusion and technology transfer, then TRIPS may have achieved some of its aims. However, the authors strike a note of caution: whether the increased trade represents increased access or merely higher prices remains an important issue for future work. Roberta Piermartini and Stela Rubínová have done seminal work for their chapter, showing how global production chains can facilitate the access to foreign non-codified knowledge, by intensifying communication between foreign firms and domestic suppliers. Patents and scientific publications are a public source of codified knowledge. However, not all knowledge can be codified and part of it diffuses only via face-to-face communication and direct transfer. To benefit from knowledge generated elsewhere, firms often need to have access also to the non-codifiable part. This boosts innovation. Using industry-level R&D and patent data for a sample of 29 countries for the period 2000–2008, the authors show that production linkages can be an important driver of knowledge spillovers, especially when the intermediate input suppliers are sufficiently integrated in the supply chain. Quantitatively, they show that international knowledge spillovers from such production chains are larger than previously estimated. Furthermore, they show that while knowledge spillovers diminish with geographical distance, distance has become less important as production chains have become more globalized.

  



Nikolas J. Zolas and Travis J. Lybbert, use an original algorithmic concordance that links patent classification to industry and product classifications to look at global value chains (GVCs) from the opposite angle to the previous chapter: how do patents affect trade and, more specifically, the organization and structure of GVCs? We know that patents play an important role in the global economy through their impact on technology diffusion, knowledge transfer and competition. The authors suggest that as production has become fragmented into coordinated processes that span the globe, the role of patents has evolved. Patents can transmit technical knowledge and facilitate upstream and downstream licensing arrangements, thereby increasing the flow of intermediate and final goods and catalyzing knowledge spillovers across sectors and regions. Using the World Input-Output Database as the basis for various GVC measures, the authors find that increased international patenting inflows are associated with greater value-added production, but that this positive effect is driven entirely by industries in high-income countries. They also find some evidence of heterogeneity in these effects by region and sectorspecific R&D intensity. They are cautious about important limitations of this exploratory analysis and implications for future research. Michael D. Smith and Rahul Telang synthesize the academic literature to help policymakers better understand the empirical evidence of both the impact of piracy on worldwide trade in copyrighted goods and the effectiveness of various strategies for protecting intellectual property from the threat of piracy. While the digital revolution has made it easier for small creators and innovators to have a larger global audience, it has also led to increased piracy of creative content. The authors find convincing empirical evidence, including from their own work, that suggests that (1) piracy harms sales in legal markets in the vast majority of settings in which it occurs, (2) lost revenue from piracy impacts producers’ incentives to innovate and create new content; more specifically, there is emerging evidence that reduced revenue from piracy is causing a reduction in both the quantity and quality of content produced by the copyright industries, (3) anti-piracy measures can be effective at changing consumer behaviour; in particular, each of the major stakeholders in content delivery – governments, right holders, platforms, and ISPs – can play a role in changing consumer behaviour. Moreover, the literature suggests that active enforcement is important to influencing consumer behaviour and that anti-piracy measures are likely to have a stronger impact on digital sales channels than on physical channels. However, the authors conclude that more empirical research, particularly



    

field experiments, are needed to replicate and refine these results to better inform policymakers. Keith Nurse and his co-authors assess the challenges and opportunities for creative industries in developing countries, especially as it relates to making the transition to a digital creative economy and to the adoption of blockchain technologies. The music industry is used as a case study because it is this sector that has experienced the greatest level of digital disruption and it is also among the first to embrace blockchain technology. The authors explore blockchain technologies as one element of the digital economy that has the capacity to present opportunities for these countries to increase market penetration and deepen global valueaddition. They identify four potential interventions that would enhance participation in the digital marketplace: promoting distribution of music on key platforms; building institutional and human resources in the creative industries to tap into the opportunities in the digital and blockchain arenas; employing targeted marketing strategies suitable for the various genres and styles of creative content on digital platforms; and strengthening the legal, technological and institutional infrastructure for collective management of copyright, through collective management organizations that stimulate creativity and musical diversity through the creation of works of local genres, generate royalties, and also undertake support activities such as education, awards and social-welfare actions.

Policy, Regulatory and Legislative Frameworks Part IV sets out an analysis of the policy, legal and regulatory frameworks that govern cross-border knowledge flows. Partly due to the intangible quality of knowledge flows, and in view of the general character of knowledge as a public good, knowledge flows are particularly sensitive to policy and legal settings, both within national jurisdictions and in the international sphere. It is, after all, the IP system itself that defines exclusive rights over knowledge products, such as inventions and technical knowhow, that would otherwise be non-excludable and nonrivalrous public goods: in other words, it is the very existence of IP laws that create and define the packages of knowledge that can be traded internationally. The efficiency and effectiveness of international knowledge flows therefore depend in part on how national regulatory and legal systems define and administer IP rights, and in part on how international rules provide for reasonable and workable interoperability between IP

  



systems that are still essentially domestic and national in character. This part of the book explores the regulatory responses to the challenges posed by new technologies, as domestic regulators and international negotiators seek to adapt and supplement existing rules to the contemporary context for knowledge flows and transactions in IP rights and digital products. Mary LaFrance examines the legal regimes that govern the streaming of music and audiovisual works in the United States and the European Union. The author concludes that although these legal systems have already undergone considerable change during the Internet era, they must continue to evolve in order to address the challenges presented by licensed and unlicensed streaming. For example, with respect to licensed streaming, the number of licenses needed is greater, and using them effectively requires tracking the utilization of each work as well as linking each work to all of the rights holders entitled to compensation. Due to their greater bargaining power, larger rights holders can typically negotiate favourable licenses, leaving independent artists on their own to protect their rights. Within the EU, consumer demand for borderless streaming has highlighted the need for multi-territorial licensing. In the case of unlicensed streaming, neither the US nor the EU have imposed filtering or monitoring obligations on ISPs that host user-provided infringing content; and the law in these jurisdictions seems to be trending in opposite directions. Paradoxically, if ISPs are required to monitor and filter their content, this is likely to favour the larger ISPs that can afford to create and invest in this technology, while creating a barrier to entry for newer services. Mira Burri discusses the depth of disruptive changes needed in trade law and policies, given the trajectory of the digital transformation of international trade. In order to do this, she examines existing regulatory frameworks – both at the international and at the regional and bilateral levels. She is, in particular, interested in an understanding of the new templates for electronic commerce and for this the chapter focuses on the most advanced model so far that of the Comprehensive and Progressive Agreement for Transpacific Partnership (CPTPP) and the United States–Mexico–Canada Agreement (USMCA). She goes on to contextualize and assess the impact of the existing legal framework, as shaped in recent times almost predominantly by preferential trade agreements. She asks whether there are better ways to address the challenges of the data-driven economy and what the essential elements of a working regulatory model should be. In conclusion, she suggests a return to the multilateral forum of the WTO and briefly sketches viable paths to do so.



    

Nigel Cory explores three interrelated topics: first, the role of data flows generally and the impact of policies that act as barriers to their movement across borders; second, how the intersection of intellectual property rights, data flows and digitalization raises new challenges for copyright enforcement and protection of trade secrets; and finally how data flows and digital trade can be indirectly affected by policies that restrict the Internet-based platforms and services that are key agents in facilitating and managing data flows, services and digital intellectual property. The author concludes that while data and digital technologies have transformed how organizations trade, especially in knowledgeintensive goods and services, many policymakers are struggling to adapt domestic and international rules and norms. This is evident in the evergrowing gap between technological innovation and domestic/international policy frameworks, which allows countries to enact policies that detract from the potential economic and social benefits of these technologies. He views intellectual property as crucial to supporting innovation and asserts that policymakers will soon face a key choice in deciding whether they want to embrace a truly global market for digitally enabled trade in goods and services and be willing to push back on the emerging threat of digital protectionism. Lucas Spadano and Luiza Coelho provide a basic understanding of how multilateral, plurilateral and bilateral trade agreements could affect cross-border knowledge flows. In particular, the authors examine whether the current multilateral disciplines are sufficient to regulate and encourage cross-border knowledge flows, or whether new trade rules to further promote such flows are needed. The question is discussed, particularly, in the light of new trends identified in recent plurilateral and bilateral agreements. The authors describe how the current international trade legal system regulates cross-border knowledge flows. The relevant WTO disciplines are explained briefly, alongside current trends in bilateral and plurilateral trade agreements. They discuss possible reasons for reforming or updating current multilateral trade rules and conclude that there is a need for such reform. Jacques de Werra and Jeff C. Dodd view knowledge-based IP transactions as vitally important to the modern global economy. The focus of their chapter is on cross-border trade in IP and information products that manifests itself through licensing. There is no internationally recognized general statement of the commercial legal principles dealing with this unique but common transactional form of licence nor on the unique but common qualities of knowledge-based intangibles. Various private

  



ordering regimes have emerged to address discrete types of transactions, whether by industry or other groups. One form of private ordering, crossborder licensing practices – which they term ‘information law merchant’, modelled on the cross-jurisdictional lex mercatoria developed by European traders in the medieval period – covers only a limited number of participants who routinely trade in certain types of information assets. Effective participation in many forms of international commercial IP transactions often requires access to specialized legal knowledge, extensive commercial experience, and elaborate contract forms, creating barriers to entry for new actors, disadvantaging smaller firms, increasing transaction and performance costs, and fostering disputes. Another form of private ordering relates to FRAND licensing across different jurisdictions. Recent developments show a piecemeal and fragmented development of FRAND licensing terms and conditions (beyond the issue of the financial remuneration or royalties), which provide no firm foundation for addressing the trade in IP assets. The authors conclude that international organizations that deal with IP issues should undertake a project to address the general principles relating to international IP commercial contracts. Peter K. Yu focuses on the legal frameworks governing the cross-border flows of machine-generated data in an era when the Internet of Things has slowly transformed into the ‘Internet of Everything’. Data generated or collected by networked sensors, interconnected devices, and intelligent machines is highly valuable. The author’s view is that if countries began to create new rights in these data – whether based on the EU proposal or other proposals – the cross-border flow of such data would raise questions about the need for new trade standards. He highlights two sets of challenges concerning the development of new trade standards for regulating the cross-border flow of machine-generated data: (1) national policy developments and (2) international norm setting. He identifies policy questions that have to be addressed before the creation of a new national regime for the protection of machine-generated data. He then turns to the potential complications that would arise in the international norm-setting arena. Taken together, these two sets of challenges show how the protection, regulation and overall governance of machine-generated data may not fit well with the existing international trade regime.

Looking Forward Part V comprises a forward-looking chapter that reviews the current status of our understanding of knowledge flows in trade, considers what we have



    

learned about the gaps in our conceptual understanding and empirical data, and considers what can be done to address these gaps. The chapter discusses the possible directions for building a stronger empirical and theoretical platform to support further policy dialogue and research. This includes improving the reach and quality of statistics, and improving the analysis of available data, including through greater integration of different forms of data (such as measurement of the value and trends of trade as a proxy for knowledge transactions, and more nuanced and granular analysis of IP statistics). Policy and legal analysis need to ‘unbundle’ the knowledge component from various proxies for it (such as trade in IP-rich goods, trade in knowledge-intensive services, and trade in IP rights for such goods as software and creative content). This would help address such shortcomings as the distorting effect of taxation arrangements and the mismatch between trade statistics and the true character of cross-border commercial transactions and the related flow of valuable knowledge. More importantly, it would provide a surer foundation for policymakers to identify and implement positive opportunities to leverage indigenous knowledge resources and incoming knowledge flows for sustainable development and welfareenhancing trade and economic policy. These developments pose challenges for policy coherence and international governance, given the diversification of approaches at the domestic level and in the emerging patchwork of bilateral and regional agreements, in contrast with the policy coherence and interoperability between jurisdictions that was brought about by the WTO TRIPS Agreement. The chapter discusses the prospects for identifying, at least in principle, a potential role for multilateral institutions to update and consolidate the governance framework for knowledge flows, or whether it is more feasible and realistic to work towards policy coherence through dialogue and cooperation, supported by a forward-looking research agenda.

Conclusions The dialogue between scholars, analysts and policymakers that led to the elaboration of this volume, and the process of editing their excellent and broad-based contributions, have reaffirmed our firm conviction of the significant value of contemporary cross-disciplinary research in this critical area, as well as its importance for policymakers seeking to grasp and identify the need and potential directions for this vital research agenda. We close with an overview of some of the more compelling areas for future work in this area.

  



Metrics. At a time when trade policy interests and settings are ever more influenced by close attention to trade balances and the impact of dispersed production and value chains, it is striking that some of the most valuable elements of international trade – those that entail transactions purely in knowledge as such – are among the most poorly measured, both in terms of specific bilateral flows and general international patterns. There are ‘known unknowns’ – such as full data on international trade for consumer licences for IP content, and on the intangible components of dispersed development and production chains – that, if addressed, could significantly alter perceptions of trade balances and trade policy priorities. The making of rules. The conventional legal paradigm for IP law and related regulations governing the knowledge economy – epitomised by the TRIPS Agreement and its implementation – is the articulation of an international, multilateral set of rules which national legislation subsequently implements. However, the emerging international framework has become ever more complex, with overlapping and potentially conflicting rules flowing from bilateral and regional agreements, the de facto reach of domestic laws of the major players and the kind of private ordering that may result in an information law merchant as suggested in this book. Regulatory impacts. In 1992, Romer observed that ‘we must recognize that ideas are economic goods which are unlike conventional private goods and that markets are inherently less successful at producing and transmitting ideas than they are with private goods’.6 This observation underscores the strong sensitivity to regulatory structures of the dissemination of usable knowledge; to some extent, it is the regulatory system – and specifically IP laws – that enable markets in ideas to develop and that structure and define those markets. Regulatory intervention in a range of areas doubtless is a key factor determining how successfully markets function to ensure that ideas are generated and are disseminated and used effectively for the overall public benefit, and regulation can distort or inhibit knowledge flows, sometimes in unexpected ways. Shifting production and innovative capacities. The mobility of knowledge and of knowledge workers, the possibility of dynamic knowledge networks not defined by national boundaries or geographical distance and the unbundling of production capacity, together with the

6

At 89.



    

evolution of a number of developing and transition economies as centres of knowledge production and technological innovation, all contribute to fast-evolving and ever‑diversifying ecosystems for innovation and the production and dissemination of knowledge. These factors are reshaping how, where and by whom innovation takes place, how and where the resultant new knowledge is developed, transacted and disseminated, and how countries can benefit directly and indirectly through technology spillovers. New ways of trading. The World Wide Web was invented by Tim Berners-Lee at CERN, in Geneva, exactly at the time that TRIPS negotiators were concluding the text of an agreement which essentially assumed that the IP component of trade largely comprised value added in traded physical goods. Since then, through the massive expansion and diversification of the Internet as a common platform for the exchange of intangible content, intellectual property rights – whether through transfer of ownership or diverse forms of licensing – have emerged as widely traded goods in themselves, both between firms and in consumer markets. The practical convergence of the ‘two IPs’ – the Internet protocol and intellectual property – since that time has opened up entirely new markets (and important non-market avenues) for the dissemination of knowledge and creative content, greatly facilitating knowledge flows and creating new possibilities for economic development and more equitable access to international markets.

A COVID-19 Epilogue This project had already been completed when the first reports emerged of a novel coronavirus that rapidly escalated into the pandemic that has since convulsed societies across the globe, caused widespread death and suffering and brought chaos to national economies and to international trade. No attempt is made here to assess or respond to the gravity and severity of this fundamental upheaval and its impact on and implications for the issues explored in this volume. Nonetheless, it is clear that the value of innovation systems and the equitable dissemination of the benefits of new knowledge production is more acutely understood than ever, and that the disruption of trade and potentially of established production chains is likely to accentuate the need and demand for more effective sharing and absorption of knowledge through digital platforms and through the licensing and other transmission of knowledge products

  



as policymakers seek first to understand and then to adapt to the ‘new normal’. The editors therefore express the hope that the ideas and possibilities explored in this volume will help to illuminate the difficult path ahead through a more complex policy landscape.

Bibliography Baldwin, Richard, The Great Convergence: Information Technology and the New Globalization, Cambridge, MA: Harvard University Press, 2016. Romer, Paul M., ‘Two Strategies for Economic Development: Using Ideas and Producing Ideas’, Proceedings of the World Bank Annual Conference on Development Economics 1992, Supplement, Washington, DC: World Bank Economic Review, 1993. Watal, Jayashree and Taubman, Antony, The Making of the TRIPS Agreement: Personal Insights from the Uruguay Round Negotiations, Geneva: WTO, 2015.

PART I Conceptual Framework

2 The Shifting Contours of Trade in Knowledge: The New ‘Trade-Related Aspects’ of Intellectual Property   Abstract This chapter charts the evolution and diversification of trade in knowledge that has taken place in the quarter-century since the WTO TRIPS Agreement came into force. Entirely new markets have come into being, potentially redefining the very character of ‘trade’. The disruptive effect of digital technology has led to much of the content – formerly conceived of as ‘added value’ embedded in physical carrier media, traded and measured as ‘goods’ – can be traded in the form of specific licences that use IP rights covering the content that is increasingly accessed online in digital form. These new forms of exchange in valuable intangible content confront fundamental assumptions about the nature of trade and its interaction with the IP system, forcing a rethink of what constitutes the ‘trade-related aspects’ of intellectual property. The issues examined include the principle of territoriality of IP rights and the segmentation of markets according to national jurisdictions; the structuring of cross-border commercial exchanges into the two discrete categories of ‘goods’ and ‘services’; the emerging disparity in regional trade agreements between provisions on digital IP standards and on digital products and e-commerce; and the significance of IP rights being treated as assets in investment treaties. Whatever formal or legal overlay is applied to these new trading arrangements – it is essential to understand that this is now trade in IP licences as such, rather than trade in goods that have an IP component as an adjunct or ancillary element. TRIPS came about at a time when economic growth theory incorporated intangible knowledge as an endogenous factor, rather than maintaining it as exogenous to models of growth. Trade policy must similarly work to incorporate an understanding of the trade in IP licences itself within cross-border commercial exchanges as an integral element of international trading relations: sale and licensing of IP rights can then be considered ‘endogenous’ to trade. This is essential for an accurate empirical picture of trade relations today, given the economic significance both of dispersed global value chains and of trade in ‘pure’ IP content as such particularly in the creative sectors.





 

Overview: The New ‘Trade-Related Aspects’ of IP Growing recognition of the economic significance of the IP system and of the knowledge component of trade led negotiators, a generation ago, to formulate the WTO TRIPS Agreement. TRIPS ostensibly addressed the ‘trade-related aspects’ of IP rights and integrated the IP system into the framework of multilateral trade law, but it was concluded before the effects were felt of the impact of digital technologies on how knowledgerich products – tangible and intangible – are traded and disseminated. In the quarter-century since TRIPS came into force, trade in knowledge – the tangible and intangible ways valuable content is disseminated through commercial channels – has dramatically evolved and diversified. Entirely new markets have come into being for valuable content defined and protected by IP, and IP licences governing knowledge constitute one form of linkage in dispersed value chains. Consequently, the ‘traderelated aspects’ of IP have evolved and diversified, to the point that trade in IP rights as such is commonplace and defines major consumer markets for creative content. The emerging trade in IP rights as such, and the digital products defined and traded through IP rights, suggests that the very character of what is conceived of as ‘trade’ needs reconsideration. For instance, from a trade perspective, valuable content (such as musical works, software or literary works) was earlier conceived as ‘added value’ embedded in physical carrier media, the physical objects that were traded and counted as ‘goods’. Trade in music, software or literature would be measured through the physical transfer of such goods, and transactions defined by the transfer of ownership over such physical objects. The disruptive effect of digital technology means that much of this content is now traded digitally. The individual transactions that make up the trade increasingly take the form not of transfer of ownership of carrier media, but of specific licences that combine contractual undertakings with limited licences to use IP rights covering the traded content. Established notions of ownership and property are under strain. Trade in content as such, valued for its own sake, therefore increasingly supersedes trade in the physical carrier media that had earlier served as a proxy for trade in valuable intangible goods such as musical and literary works, software and games. These new markets and new forms of exchange in valuable intangible content confront fundamental assumptions about the nature of trade and its interaction with the IP system, forcing a rethink of what constitutes the ‘trade-related aspects’ of IP.

                                 



This dynamic context, and the disruptive effect of technological change, have implications for the principle of territoriality of IP rights and the segmentation of IP markets according to national jurisdiction; the structuring of cross-border commercial exchanges into the two discrete categories of ‘goods’ and ‘services’; the emerging disparity in regional trade agreements between digital IP standards and provisions on digital products and e-commerce; and the significance of IP rights also being treated as assets in investment treaties. Whatever formal or legal overlay is applied to these new trading arrangements, it is essential to build an understanding of their character, in part, as trade in IP licences as such, rather than viewing the IP component as an adjunct, ancillary or extraneous element. Just at the time TRIPS was being negotiated, Romer and others demonstrated the need for economic growth theory to incorporate intangible knowledge as an endogenous factor. Today, trade policy must similarly work to incorporate an understanding of the IP dimension of cross-border commercial exchanges as an integral element of trading relations – and ideally count them more comprehensively and effectively in the trade balance statistics that can influence trade policy choices. This means treating the exchange and licensing of IP rights systematically and effectively as ‘endogenous’ to trade. This shift in framing the ‘trade-related aspects’ of IP – indeed, considering the IP-related aspects of trade – is essential for an accurate empirical picture of trade relations today, given the economic significance both of dispersed global value chains and of trade in ‘pure’ IP content as such, particularly in the creative sectors. Such a move need not challenge formal positions as to the categorization of trade in digital products. Instead, it may be more productive to accept the diverse range of interactions between the IP system and trade, and build an understanding on empirical observation of diverse phenomena, rather than a priori formal categorization.

‘Old Town Road’ and New Trade Routes The song ‘Old Town Road’, by the rapper Lil Nas X (Montero Hill), set new sales records in 2019, even as it resisted definition within the music business’s established categories. For an unprecedented 19 weeks, the song topped the Billboard Hot 100 , the principal music industry listing of currently popular recordings in the United States – the original Billboard magazine ‘hit parade’ dating back to the 1930s. The Recording Industry



 

Association of America (RIAA) later certified the song as ‘diamond’, representing confirmed sales of 10 million units; it reached this milestone in record time.1 And the song defied conventional musical genres: blending elements of hip hop, trap and country music. ‘Old Town Road’ was initially listed on Billboard’s Hot 100 and on its R&B/hip hop chart, as well as on its country music chart, but was reportedly removed from the latter on the grounds that it ‘does not embrace enough elements of today’s country music to chart in its current version’, a decision that sparked controversy, it remained on the rap charts.2 At the time, the songwriter argued that the ‘song is country trap. It’s not one, it’s not the other. It’s both. It should be on both [charts]’.3 The song was built on the foundation of a beat created by Kiowa Roukema (under the handle YoungKio), a Dutch musician. He had encountered – when browsing YouTube algorithmic recommendations – the instrumental track ‘34 Ghosts IV’ produced by Nine Inch Nails (composed by Trent Reznor and Atticus Ross). He combined a trap rhythm with a banjo riff sampled from the Nine Inch Nails track. In turn, Mr Hill obtained this beat from BeatStars, an online market for instrumental audio clips to be used by recording artists and songwriters. While he was reported to have ‘purchased’ or ‘bought’ the beat, in fact he took out a standard, non-exclusive licence – a ‘basic lease’ for USD 30, which gave limited rights to use the clip for recording music that was capped by a permitted number of copies, and limited rights to audio streams and broadcasting rights of the resulting musical work. BeatStars is an international marketplace for intellectual property licences, the territory of application of which is defined as ‘the world’. It reportedly hosts over 1.5 million beats, audio clips available for licensing, that are posted by music producers from around 160 countries, and has paid over USD 50 million to producers such as Mr Roukema – although it is very unlikely that these payments will be counted as exports of goods or services from the countries in which they work. The Nine Inch Nails track had been released under a Creative Commons licence, with a positive encouragement for the material to be used for non-commercial

1

2

3

Rania Aniftos, ‘Lil Nas X’s “Old Town Road” Is the Fastest Song in History to Be Certified Diamond by the RIAA’, Billboard, 22 October 2019. Hubert Adjei-Kontoh, ‘Lil Nas’ song was removed from Billboard for not being “country” enough. But who gets to decide categories?’, The Guardian, 2 April 2019. Andrew Chow, ‘Lil Nas X Talks “Old Town Road” and the Billboard Controversy’, available at https://time.com/5561466/lil-nas-x-old-town-road-billboard/.

                                  



purposes. However, the right to use the sample in ‘Old Town Road’ was only cleared retrospectively through a telephone call. The original musical composition was credited to Hill and Roukema as well as to Trent Reznor and Atticus Ross: of the four credited songwriters, only the last two had actually met, and the first two were separated by an ocean. Mr Hill initially released the song independently (signing to an established record label, Columbia, only after its breakout success) and promoted it through extensive use of social media, including the newly popular avenue of TikTok. He encouraged its use in video clips by other users, and gained attention to the track through a variety of Internet memes. The groundswell of interest thus generated led to its popularity in more conventional settings, in particular those used by Billboard and other industry sources to measure success. The Billboard Hot 100 measure includes traditional retail sales of physical media (such as CDs), as well as digital downloads, streaming and on-demand access.4 ‘Old Town Road’ exemplifies how digital technology has transformed the music industry in particular in the period since the WTO TRIPS Agreement was concluded in 1994, some five years before the song’s creator and performer was born. This transformation is evident in terms of how the song was composed and disseminated, the interplay between creators’ rights and user rights as its components came together, the geographically dispersed cluster of rights over the song’s ingredients (the very process of its composition and production resembling a global value chain), the significance of digital platforms including a global marketplace for IP licences unsupported by physical carrier media, the song’s non-conformity with established genres, and even the manner of measuring its success, all taken together, offer illuminating insights into the impact of digital disruption on this sector. It demonstrates how entirely new forms of trading in creative content are supplanting traditional business models and challenging traditional genres. ‘Old Town Road’ exemplifies trends that can be said to be transforming the very nature of trade, and the linkages between the intellectual property (IP) system and international trade: • the development of online commercial markets for licences in IP content as such, enabling seamless trading in a global space, and transforming the notion of ‘trade-related aspects’ of IP rights;

4

See www.billboard.com/p/billboard-charts-legend.



 

• the diversification and informality of the means of creation, dissemination, reuse and consumption of digital content; • the multiple uses of the Internet as a packet-switched network, and the transmission of packets of data structured by the Internet protocol (TCP/IP), for diverse creative, cultural, social and commercial purposes, and the correspondingly diverse legal statuses of those packets of data, which may variously be construed as in the public domain, subject to fair use or user rights, or bound by contractual obligations and IP rights; • the blurring of boundaries: between producer and user of creative content, between established genres, between distinct distribution channels, and between the distinct national jurisdictions under which IP rights are recognized and enforced; and • the challenge of tracking and measuring international commercial transactions that entail the licensing of IP content no longer embedded in the physical carrier media – discs, tapes, printed publications such as sheet music – that once could comprise the totality of creative content traded internationally.

From Trade in Atoms … GATT and the ‘Trade in Goods’ Paradigm The multilateral trade agreements that came into force on the establishment of the WTO in 1995 were negotiated and concluded in an effectively pre-Internet age, with scant consideration of how digital technologies were likely to reshape not only the directions and composition of trade, but even transform the very legal and economic character of commercial transactions that constitute trade. The interaction between trade and the IP system, and the general acceptance of the significance of the intangible component of trade, had only slowly taken root in a multilateral system of trade rules that originally focused on trade in physical goods – informally, ‘things you could drop on your foot’. In 1947, the General Agreement on Tariffs and Trade (the GATT) established a rule, essentially, of non-interference and of non-discrimination between trade law and IP law: the IP system was held at arm’s length. Thus, GATT Article XX, on general exceptions, accepted that contracting parties may choose to protect IP within their domestic systems, without requiring this. At the same time, it provided that IP protection should not be a disguised restriction on trade, and

                                  



should not discriminate against imported goods (the GATT being a trade-in-goods agreement, discrimination was construed in terms of goods and not of persons): in effect, protect IP if you wish, but do no harm to trade in physical goods. GATT jurisprudence confirmed that measures to enforce IP at the border could fall foul of the principle of non-discrimination, effectively setting limits on the impact of IP measures on trade in goods.5 One strand of IP – loosely, the unfair competition/geographical indication (GI) nexus – found some, tentative, recognition in the original GATT text. The obligation under Article IX.6 to cooperate to prevent the use of trade names that would ‘misrepresent the true origin of a product, to the detriment of such distinctive regional or geographical names of products of the territory of a contracting party as are protected by its legislation’ recalls the trade-relatedness of geographical indications and the mechanism for communicating names protected under domestic law points towards a later mandate to negotiate a multilateral GI register, but there are no obligations positively to protect such terms. Over time, trade policy began to register the relevance of IP issues more widely. Thus a GATT inventory of non-tariff barriers to trade compiled in 1968 included a notification by the United Kingdom of Italy’s local working requirements for patents (invoking the Paris Convention for the Protection of Industrial Property), and of the manufacturing clause in US copyright law.6 A more sustained and structural shift to recognizing the intangible value that IP embeds in traded goods can be traced more concretely to past GATT work on counterfeit trade beginning in the 1970s. The United States sought to introduce a requirement to act against trade in counterfeit goods within the GATT Tokyo Round negotiations: an early proposal was headed ‘Agreement on the Sanctions to be Imposed Upon the Importation of Counterfeit Merchandise’,7 and dealt with international trade in articles that bore a spurious trademark or were made in violation of copyright. A subsequent version, submitted jointly with Canada, the European Community and Japan, proposed an ‘Agreement

5

6

7

See United States – Section 337 of the Tariff Act of 1930, Report by the Panel adopted on 7 November 1989, L/6439 – 36S/345. GATT documents COM.IND/4 (30 August 1968) and COM.IND/4/Corr.1 (26 September 1968). GATT document MTN/NTM/W/204, Multilateral Trade Negotiations – Group ‘NonTariff Measures’ – Sub-Group ‘Customs Matters’ – Commercial Counterfeiting – 11 December 1978.



 

on Measures to Discourage the Importation of Counterfeit Goods’,8 which only addressed international trade in imported goods bearing a false representation of a trademark. These proposals were unsuccessful in that they failed to produce an immediate negotiating outcome, but they helped shape the negotiating mandate for the subsequent Uruguay Round negotiations – the Punta del Este Declaration of 20 September 1986. This negotiating mandate – within the broader category of ‘trade in goods’ – referred to ‘[t]rade-related aspects of intellectual property rights, including trade in counterfeit goods’ and established the negotiating objective of developing ‘a multilateral framework of principles, rules and disciplines dealing with international trade in counterfeit goods’. Hence, for historical and practical reasons, the essential notion of IP and international trade that the TRIPS mandate addressed was still centred on the idea that IP was embedded in physical goods, and that the infringing trade to be suppressed was trade in infringing goods. This practical assumption that international trade relating to IP was embedded in physical goods finds a strong echo in the earlier multilateral conventions on IP administered by the World Intellectual Property Organization (WIPO). Thus, the Paris Convention deals with ‘articles’, ‘products’, and ‘goods’ as either embodiments or infringements of IP. Until its revision in 1958, Paris had not recognized service marks at all, in line with the past understanding in many countries that trademarks were reserved for goods only. What was arguably the first truly ‘trade-related’ agreement on intellectual property – the Madrid Agreement for the Repression of False or Deceptive Indications of Source on Goods of 1891 – squarely addressed trade in goods in a manner that has consonance with the one positive obligation related to IP in the original GATT, the requirement under Article IX.6 to cooperate ‘with a view to preventing the use of trade names in such manner as to misrepresent the true origin of a product’. In this ‘analogue’ age, it was, indeed, the case that goods served as proxies especially for consumer access to the content protected and defined by IP rights. Hence, for most consumers, music, books, journals, cinematic works, and even consumer software and games, were traded internationally almost exclusively on physical carrier media – optical and vinyl discs, recorded tape, film stock, newsprint, bound volumes, sheet music and so on. And guidelines for measuring international trade in 8

GATT document L/5382, Agreement on Measures to Discourage the Importation of Counterfeit Goods, 18 October 1982.

                                  



goods recognize such material as traded goods: ‘[a]s a general guideline, media, whether or not recorded, is included in international merchandise trade statistics at its full transaction value, except for media used for carrying customized software or software written for a specific client or originals of any nature, which should in principle be excluded.’9 One exception concerns international ‘newspapers and periodicals sent under direct subscription’10 – though they entailed the passage of valuable physical goods across borders, these were counted as trade in services, the physical shipment of the publications being conceptually subordinate to the provision of a subscription service. Commercial transactions defined by the exchange of ownership of these physical things – these bundles of atoms – therefore served as proxies for the acquisition of access to valued intangible content (such as music, literature, films and software). The reliance on physical carrier media as the means of dissemination meant that international trade in musical and literary works could readily be tracked and relatively accurate trade statistics maintained, by monitoring the quantity and value of physical media that passed through customs procedures on a country’s boundary. Hence recorded media (recorded discs, tapes and film stock, published books, newspapers) have been counted and valued separately from blank media (blank discs, tape and film stock, paper, newsprint); their ‘full transactional value’ incorporates the licence acquired to use the recorded content, normally limited to certain private uses. Broadcasting and other forms of wired and wireless transmission – on domestic and cross-border networks – also provided the public with access to some categories of material, but not in a form susceptible to individual transactions concerning ownership of a right to continued access to a specific work.11 In categorizing trade in goods for customs valuation and statistics, the Harmonized Commodity Description and Coding Systems (HS) classification of traded goods distinguishes between recorded and unrecorded optical media, magnetic media and photographic media. Clear differences in value are apparent depending on whether blank or recorded media are traded. Thus, in 2018, a global total of 13.8 billion imports of

9

10 11

United Nations, International Merchandise Trade Statistics: Concepts and Definitions 2010, New York: IMTS, 2010, p. 15. Ibid., p. 21. The lack of such a right was a factor behind the negotiation of what were termed the ‘WIPO Internet treaties’ in 1996, notably producing a new right of distribution – see ‘IP in the Digital Marketplace: The Effect of Exhaustion’, infra.



  2018 exporters of product 852341: optical media for the recording of sound or of other phenomena, unrecorded USD mn

300 250 200 150 100 50 0

Figure 2.1

List of exporters in recorded optical discs in 2018

Source: ITC Trade Map.

recorded optical discs (HS code 852349) was reported, as against 2.1 billion imports of unrecorded (blank) optical discs (HS code 852341). The contrasting two patterns of exports (see Figures 2.1 and 2.2) illustrate a kind of ‘value chain’ in international trade, as intangible content is added to the blank discs, greatly enhancing their value, with the export profile of the more valuable recorded media being dominated by developed countries.

Historical Legal Questions: What Are You Purchasing? Prefiguring the manner in which we need today to consider the implications of digital disruption for trade in knowledge products such as musical works, earlier technological changes – such as the development of soundrecording and broadcasting technologies – led to similar issues concerning the very character of trade in copyright works. Hence, this question arose at a time when radio technology made it possible for a sound recording – then normally distributed on gramophone records – to be widely broadcast to the public beyond the original, private, context in which such records had been played. Until the advent of broadcasting, the only ways of gaining access to a musical performance were either being physically present at the performance or by means of access to a physical object, the recorded disc. Hence the question arose, with resonance still today: if you do buy a recorded physical carrier medium, such

                                 

2250



2018 exporters of product 852349: optical media for the recording of sound or of other phenomena (excluding unrecorded ... ) USD mn

2000 1750 1500 1250 1000 750 500 250 0

Figure 2.2 List of exporters in unrecorded optical discs in 2018 Source: ITC Trade Map.

as a gramophone record or an optical disc, what are you actually purchasing? What is the very nature of that transaction? Is it essentially the purchase of a physical thing, a ‘chattel’ in the legal jargon, something that you have property rights over and can sell and pass on to others? Is it the purchase of a rather limited private right to use the content, to use the intangible content? Is it a combination of the two? And how is the nature of this transaction affected by the introduction of new technologies for disseminating content? In an earlier age, when radio broadcasting was a disruptive new technology, a US court developed an intriguing way of answering these questions. In the 1930s, the musician Fred Waring, bandleader of the Pennsylvanians, ran into difficulty enforcing an agreement with his record company, the Victor Talking Machine Company. His band’s recorded music was licensed for gramophone production provided the ensuing records were labelled ‘not licensed for radio broadcast’. The broadcast market was a separate, valuable source of income. The radio station WDAS purchased the disc and broadcast these recordings contrary to the asserted licence conditions. This led the court to construe the significance of technological disruption for rights and entitlements linked to intangible content: The problems involved in this case have never before been presented to an American or an English court. They challenge the vaunted genius of the law to adapt itself to new social and industrial conditions and to the



  progress of science and invention. For the first time in history human action can be photographed and visually re-portrayed by the motion picture. Sound can now be mechanically captured and reproduced not only by means of the phonograph for an audience physically present, but, through broadcasting, for practically all the world as simultaneous auditors. Just as the birth of the printing press made it necessary for equity to inaugurate a protection for literary and intellectual property, so these latter-day inventions make demands upon the creative and ever-evolving energy of equity to extend that protection so as adequately to do justice under current conditions of life.12

In this particular case, the challenge of technological disruption led the court to conclude that the licence restricting the use of the gramophone record was an ‘equitable servitude’13 on a chattel. The physical disc was sold by the gramophone company to a radio station. Waring himself was not a party to this transaction, and could not interfere with the sale of the disc as such. Even so, he was entitled as a matter of equity to enforce his interest in the recorded content contained on the disc and could restrain certain uses of the disc, despite the separate transfer of ownership of the disc as personal property (the ‘chattel’). [No] valid reason exists why the restriction attached to the manufacture and sale of the records in this case should not be enforced in equity . . . in a sense the plaintiff was not imposing a restriction in connection with a sale by him of a chattel. The chattel here consisted of the phonograph record. This the plaintiff never owned. What he granted was merely the incorporeal privilege of reproducing the rendition of the song indented upon the chattel sold by the Talking Machine Co. The reservation or restriction imposed by him was to limit the extent of this privilege. The title to the physical substance and the right to the use of literary or artistic property which may be printed upon or embodied in it are entirely distinct and independent of each other.14

The court then upheld the complaint on the basis that the radio station’s use of the recording was an act of unfair competition. This somewhat experimental construction of the law did not find more general or lasting acceptance, and the law evolved to recognize a growing bundle of distinct IP rights applicable to recordings and performances as such.15 The case is

12 13 14 15

Fred Waring v. WDAS Station Inc. 194 A. 631 (Pa. 1937) 433. Zechariah Chafee, ‘Equitable Servitudes on Chattels’ (1928), Harvard Law Review 41: 945. Waring v. WDAS, p. 447. The need for ideas of equity and balance to adjust to such technological impact is discussed in Antony Taubman, ‘Nobility of Interpretation: Equity, Retrospectivity, and

                                 



cited here, nonetheless, as an instructive historical insight into the continuing challenges, in a changing technological environment, of applying legal principles that seem better adapted to deal with property interests in tangible goods, as against claims or interests relating to intangible content. Defining and articulating the limits of such claims over embedded content when distributed by means of the sale of a physical thing, a chattel, remains a task for the policymaker or the courts confronted with today’s new technological possibilities.

Exhaustion of IP Rights and the Scope of Markets Indeed, considering the trade dimension in particular, in the contemporary commercial and legal environment, the relationship between the physical carrier medium and the rights governing the content it carries has increasing significance for determining the extent to which goods carrying IP-protected content can be traded, domestically and across borders. Key to the creation and structure of such borderless markets for IP content is the question of exhaustion of IP rights, particularly copyright. The term ‘exhaustion’ refers to the generally accepted principle in IP law that a right owner’s exclusive right to control the distribution of a protected item lapses after the first act of distribution, when the right holder is assumed to have received a fair commercial return. On an international level, given the generally jurisdictionally defined and bound character of IPRs, it refers to the extent to which distinct authorization is required for IP-protected items to cross borders and to pass to distinct IP jurisdictions. The scope of exhaustion of rights forms the central pivot of the interaction between domestic jurisdictions and global markets for products defined or governed by IP rights. The TRIPS negotiators were unable to bridge between contrasting views that exhaustion should apply nationally (meaning that an IPprotected product can be further sold within the same jurisdiction, but can not be imported even if legitimately acquired with the IP owner’s consent in a foreign market), or internationally (meaning that the IPprotected item, once sold, can be imported into other jurisdictions even if

Collectivity in Implementing New Norms for Performers’ Rights’ (2005), Journal of Intellectual Property Law 12: 351, available at https://digitalcommons.law.uga.edu/jipl/ vol12/iss2/2.



 

the IP right is separately in force there).16 Thus this question was left open under TRIPS for national authorities to regulate (provided nondiscrimination is respected), a matter confirmed in the 2001 Doha Declaration on the TRIPS Agreement and Public Health.17 The central question, considering the policy dimension of exhaustion, is the extent to which the existence of jurisdictional boundaries constrains the free flow of IP-protected material, and to what extent the holder of an IP right should be able to bar its flow across borders. The approach taken to exhaustion of IP rights embedded in tangible goods will determine whether there can be, effectively, a single global market for physical goods carrying IP-protected material, superseding distinct domestic markets defined by IP rights granted under national legal systems. In turn, the contours of such markets can depend on the character of the underlying transaction when a physical good is purchased that carries content protected by IP rights. Two recent US Supreme Court decisions have considered this question when determining the effective reach of IPRs across jurisdictions when goods are traded internationally, and thus the degree to which IPRs can shape international trade flows for knowledge content defined and protected as IP. The international dimension of the Kirtsaeng18 case concerned whether textbooks lawfully sold abroad with the consent of the copyright owner (at a lower price than the domestic versions) could be imported and sold in the US. The domestic copyright law of the US provides that ‘the owner of a particular copy or phonorecord lawfully made under this title, or any person authorized by such owner, is entitled, without the authority of the copyright owner, to sell or otherwise dispose of the possession of that copy or phonorecord’.19 This provision codifies the ‘first sale’ doctrine under US law, under which a

16

17

18 19

TRIPS Agreement, Article 6. For an account on the negotiation of this issue, see David Fitzpatrick, ‘Negotiating for Hong Kong’, in Jayashree Watal and Antony Taubman, The Making of TRIPS, Geneva: WTO, 2015, pp. 285–291. Para. 5 of the Doha Declaration on the TRIPS Agreement and Public Health confirmed that the ‘effect of the [TRIPS] provisions . . . relevant to the exhaustion of intellectual property rights is to leave each member free to establish its own regime for such exhaustion without challenge, subject to the MFN and national treatment provisions of Articles 3 and 4’. Kirtsaeng v. John Wiley & Sons, Inc., 568 US 519 (2013). 17 US Code §109 (a).

                                  



copyright holder is not entitled to restrain further sales once a first sale has been made. The international aspect of the case therefore hinged on whether the phrase ‘lawfully made under this title’ was limited in reference to copies made within the US, or was not geographically restricted; the majority held that it was ‘nongeographical’, and that the domestic statute defined a standard of lawfulness, while not requiring a copy to be made in the United States to benefit from this exhaustion of the distribution right. A majority found that physical books containing copyright works, imported into the United States in the so-called ‘grey market’, were not infringements of US copyright law and could thus legitimately enter the domestic market. A host of policy questions lay behind this seemingly narrow question. Notably, from the point of view of the interplay between IPRs and trade law and policy, the majority held that the copyright statute aimed at a principle of ‘equal treatment’ between foreign and domestically produced works: ‘the “equal treatment” principle is difficult to square with a geographical interpretation that would grant an American copyright holder permanent control over the American distribution chain in respect to copies printed abroad but not those printed in America.’20 The majority decision draws an analogy between the purchaser’s rights over the imported copyright-protected textbook, and ownership of a ‘chattel’ (a tangible item of property), rather than viewing the physical book as being essentially a carrier medium for protected content and licensed IP and thus giving primacy to the IP dimension of the original transaction. In the Waring case noted above, a copyright restriction on radio broadcast of a sound recording was construed as an ‘equitable servitude on a chattel’.21 Such a construction sees the essential nature of the transaction as transfer of ownership of the physical object, while enjoyment of the possession of a physical copy of the recording was subject to continuing liability to the originator of the sound recording.22 Dissenting, Justice Ginsberg argued that ‘lawfully made under this title’ refers to ‘instances in which a copy’s creation is governed by, and conducted in compliance with, Title 17 of the US Code’. She argues 20 21 22

Kirtsaeng, n. 18 supra, 3. Waring v. WDAS Broad. Station Inc., 194 A. 631, 638 (1937). See Antony Taubman, ‘TRIPS Encounters the Internet: An Analogue Treaty in a Digital Age, or the First Trade 2.0 Agreement?’, in Mira Burri and Thomas Cottier (eds), Trade Governance in the Digital Age, Cambridge: Cambridge University Press, 2015, p. 314.



 

that, since US copyright law is not extraterritorial, it is anomalous ‘to speak of particular conduct as “lawful” under an inapplicable law’. Her dissent sets out the essential policy question as follows: Because economic conditions and demand for particular goods vary across the globe, copyright owners have a financial incentive to charge different prices for copies of their works in different geographic regions. Their ability to engage in such price discrimination, however, is undermined if arbitrageurs are permitted to import copies from low-price regions and sell them in high-price regions.23

At the international level, she outlined the divergent positions as follows: In the absence of agreement at the international level, each country has been left to choose for itself the exhaustion framework it will follow. One option is a national-exhaustion regime, under which a copyright owner’s right to control distribution of a particular copy is exhausted only within the country in which the copy is sold . . . Another option is a rule of international exhaustion, under which the authorized distribution of a particular copy anywhere in the world exhausts the copyright owner’s distribution right everywhere with respect to that copy . . . The European Union has adopted the intermediate approach of regional exhaustion, under which the sale of a copy anywhere within the European Economic Area exhausts the copyright owner’s distribution right throughout that region . . . Section 602(a)(1), in my view, ties the United States to a national-exhaustion framework. The Court’s decision, in contrast, places the United States solidly in the internationalexhaustion camp. Strong arguments have been made both in favor of, and in opposition to, international exhaustion . . . International exhaustion subjects copyright-protected goods to competition from lower priced imports and, to that extent, benefits consumers. Correspondingly, copyright owners profit from a national-exhaustion regime, which also enlarges the monetary incentive to create new copyrightable works.24

A subsequent Supreme Court decision, Impression Products,25 concerning the aftermarket for used printer cartridges, addressed the question of international exhaustion of patent rights in a case that had some similarities with (and referred to) Kirtsaeng. The international dimension of the case pivoted on whether patent-protected cartridges sold abroad could be

23 24 25

Kirtsaeng, p. 43. Ibid. Impression Products, Inc. v. Lexmark International Inc., 581 US 1523 (2017).

                                  



refilled and then sold in the United States, and the domestic dimension concerning whether conditions imposed on the subsequent use by the purchaser of cartridges could be enforced under patent rights. In both circumstances, the majority of the Court held that a sale of a patented product (even in a foreign jurisdiction) by the patent holder exhausted the patent rights. Thus the patent could not be used to prevent downstream resale or importation of a refilled cartridge. The majority decision revisited the policy considerations it had discussed in Kirtsaeng, in particular the common law ‘refusal to permit restraints on the alienation of chattels’. In the domestic market, it found that, In sum, patent exhaustion is uniform and automatic. Once a patentee decides to sell—whether on its own or through a licensee—that sale exhausts its patent rights, regardless of any post-sale restrictions the patentee purports to impose, either directly or through a license.26

And when the patented article is sold internationally and then imported to the United States, the outcome was the same: An authorized sale outside the United States, just as one within the United States, exhausts all rights under the Patent Act . . . exhaustion occurs because, in a sale, the patentee elects to give up title to an item in exchange for payment. Allowing patent rights to stick remora-like to that item as it flows through the market would violate the principle against restraints on alienation. Exhaustion does not depend on whether the patentee receives a premium for selling in the United States, or the type of rights that buyers expect to receive. As a result, restrictions and location are irrelevant; what matters is the patentee’s decision to make a sale.27

Again, Justice Ginsberg dissented from the majority finding on the international dimension , arguing that the overseas sale operated independently of the US patent system: ‘US patent protection accompanies none of a US patentee’s sale abroad—a competitor could sell the same patented product abroad with no US-patent-law consequence. Accordingly, the foreign sale should not diminish the protections of US law in the United States.’28 In the Court’s analysis, it is the act of selling an article that exhausts the patent right, and the fundamental aversion to restraints on the use of

26 27 28

Ibid., p. 1535. Ibid., p. 1538. Ibid., p. 1539.



 

physical chattels once sold forms the legal foundation for exhaustion – the exhaustion rule ‘marks the point where patent rights yield to the common law principle against restraints on alienation’. Hence it is the transfer of rights over physical property – and the consequent common law right to use and resell it without constraint from the ‘remora-like’ patent rights – that characterizes the essential transaction, not the fact that it is a sale undertaken within the panoply of applicable patent rights – in fact, the sale means that the product ‘is no longer within the limits of the [patent] monopoly’ and instead becomes the ‘private, individual property’ of the purchaser. But, again, the analysis may differ in cases where there is no tangible product, the transfer of ownership of which triggers the exhaustion of rights. Instead, the transaction may potentially be construed as the purchase of a limited licence, essentially defined in terms of a limited entitlement to use the patented invention within the scope of the exclusivity granted to the patent holder (the contrasting case of exhaustion of rights over intangible products is considered in ‘IP in the Digital Marketplace: The Effect of Exhaustion’). In this context, it is noteworthy that the majority decision in Impression Products takes care to distinguish the nature of the licence from the sale of a physical product. A patentee can impose restrictions on licensees because a license does not implicate the same concerns about restraints on alienation as a sale. Patent exhaustion reflects the principle that, when an item passes into commerce, it should not be shaded by a legal cloud on title as it moves through the marketplace. But a license is not about passing title to a product, it is about changing the contours of the patentee’s monopoly: The patentee agrees not to exclude a licensee from making or selling the patented invention, expanding the club of authorized producers and sellers. See General Elec. Co., 272 U. S., at 489–490. Because the patentee is exchanging rights, not goods, it is free to relinquish only a portion of its bundle of patent protections. A patentee’s authority to limit licensees does not, as the Federal Circuit thought, mean that patentees can use licenses to impose post-sale restrictions on purchasers that are enforceable through the patent laws. So long as a licensee complies with the license when selling an item, the patentee has, in effect, authorized the sale. That licensee’s sale is treated, for purposes of patent exhaustion, as if the patentee made the sale itself. The result: The sale exhausts the patentee’s rights in that item . . . A license may require the licensee to impose a restriction on purchasers, like the license limiting the computer manufacturer to selling for non-commercial use by individuals. But if the licensee does so—by, perhaps, having each customer sign a contract promising not to use the computers in business—the sale nonetheless exhausts all patent

                                  



rights in the item sold . . . The purchasers might not comply with the restriction, but the only recourse for the licensee is through contract law, just as if the patentee itself sold the item with a restriction.29

Already, therefore, for trade in goods, it is impossible to understand the full range of their legitimate tradability internationally – the extent to which there is a downstream market – without careful analysis of the applicability and the effective scope of IP rights covering certain goods and the circumstances in which they are considered no longer to bind a purchaser (the exhaustion question). These cases draw on a legal tradition against restraints on the alienation of chattels (or constraints on dealing with a physical good once legitimately purchased); in both the cases just discussed, the majority drew on the roots of the common law, citing the seventeenth-century jurist Lord Coke on this question: [If] a man be possessed of . . . a horse, or of any other chattell . . . and give or sell his whole interest . . . therein upon condition that the Donee or Vendee shall not alien[ate] the same, the [condition] is voi[d], because his whole interest . . . is out of him, so as he hath no possibilit[y] of a Reverter, and it is against Trade and Traffi[c], and bargaining and contracting betwee[n] man and man: and [*539] it is within the reason of our Author that it should ouster him of all power given to him.30

Pivotal, for these recent decisions, was a longstanding principle that favoured freedom to trade in legitimately purchased and owned physical goods, as against an approach that would favour the claims of the holder of intangible IP rights, rights seen perhaps as more contingent, less concrete. These cases illustrate how technological change and the growing value given to intangible content when carried on physical carrier media continues to raise a tangle of legal and policy questions. As new technologies fundamentally alter how creative works are distributed and made available to the public, established legal frameworks are put into tension. In a forward-looking mood in the 1930s, the Waring court had proclaimed that ‘there is no reason . . . why an ancient generalization of law should be held invariably to apply to cases in which modern conditions of commerce and industry and the nature of new scientific inventions make restrictions highly desirable. Mere aphorisms should not be permitted to fetter the law in furthering proper social and economic purposes.’31 29 30

31

Ibid., p. 1535. Edward Coke, Institutes of the Laws of England: Containing the Exposition of Many Ancient and Other Statutes, 1628, p. 223. Waring, n. 12 supra.



 

Similar challenges for the adaptability of the law arise today as digital technology radically transforms the way valuable intangible content is traded and disseminated, to the extent that the physical carrier medium – the ‘chattel’ – can be dispensed with altogether. The advent of trading of IP-protected materials as intangible content on digital platforms has raised the issue of whether the answer to this question can and should be different depending on whether IP is embedded in a physical carrier medium (such as a book or optical disc) or simply traded as a licensed copy of a digital file. Digital disruption has raised the parallel question over the extent of downstream control over commercial reuse that an IPR can and should confer: in particular, whether and when rights over digital content are exhausted, limiting the right holder’s entitlement to restrain further distribution of the protected digital work – in short, can legitimately purchased digital copies be resold (bearing in mind that the nature of the ‘purchase’ is essentially a limited IP licence)? One of the central trade policy questions in considering the implications of digital disruption is whether an intangible product, with similar properties for the consumer as its tangible counterpart, should be treated in the same way, or be treated differently, on the basis that it is essentially a bundle of intangible rights, with no physical substrate that can be ‘owned’ as a chattel. Would the Supreme Court have ruled differently if the textbooks concerned were e-books, and no tangible product was imported? Would both national exhaustion and digital exhaustion apply in such a case? The digital dimension of exhaustion is discussed in ‘IP in the Digital Marketplace: The Effect of Exhaustion’, below.

… To Trade in Bits The Internet Protocol as a Trade Pact? In 1990, as TRIPS negotiators progressed towards a final text, 0.05 per cent of the world’s population used the Internet; in 2018, the ITU reported that the Internet was accessible by over 51 per cent of a population that had since grown by over 2 billion. Thus, following the rapid recent uptake of the Internet in the developing world – and despite enormous disparities and troubling inequities in access – for the first time most people on the planet have access to the Internet. The emergence of Internet connectivity – from its roots in ARPANet, a US Government defence project, through a narrow band of academics

                                  



and research scientists, to reach a global public – has already, in myriad ways, transformed social relations, political dynamics and cultural life; it has opened up new opportunities for economic and social development, to the extent that Internet access is now recognized as a measure of progress towards attainment of the UN Sustainable Development Goals. One TRIPS negotiator has since observed that ‘the Internet was not then upon us [and the] negotiators did not indulge in futurology’.32 The Chair of the Uruguay Round Negotiating Group on TRIPS recalls that the negotiations proceeded unaware of the contemporaneous invention of the World Wide Web at CERN, nearby in Geneva.33 It was around the time of formal adoption of the TRIPS text, in 1994, that the implications of global connectivity through the Internet began to enter mainstream consciousness, as its use rapidly broadened beyond the research and academic communities. It was in this period that consumers (initially in a handful of developed countries, later more equitably distributed) were first gaining access to the Internet and the World Wide Web in particular, and the Internet was gaining wider practical use in many countries as a means of cultural, social and commercial exchange. When the TRIPS negotiations commenced, there were 2,308 Internet hosts in existence, primarily research and academic institutions,34 and the user base was narrow and almost entirely comprised of academics and scientific researchers. By the time TRIPS entered into force, the number of Internet hosts had grown to 5,846,000, and the growth has proceeded exponentially since then, exceeding 1 billion by 2014. The user base grew still more sharply, from roughly 40 million in 1995 to over half the world’s population by 2018. Reflecting on the transformative impact of digital connectivity, in a commentary serendipitously published on 1 January 1995, the very day the TRIPS Agreement entered into force, Nicholas Negroponte predicted a fundamental transformation in the way we communicate, transact business and share content, summarized simply as ‘bits, not atoms’. As he described this phenomenon, it had fundamental implications for the

32

33

34

David Fitzpatrick, ‘Negotiating for Hong Kong’, in Watal and Taubman, The Making of TRIPS, pp. 285–291. Lars Anell, ‘Keynote speech at the TRIPS Symposium, 26 February 2015’, in Watal and Taubman, The Making of TRIPS, pp. 285–291. See www.isc.org/solutions/survey/history.



 

way in which we assess the value of goods, and the way that we conduct transactions in goods through trade: When returning from abroad, you must complete a customs declaration form. But have you ever declared the value of the bits you acquired while traveling? Have customs officers inquired whether you have a diskette that is worth hundreds of thousands of dollars? No. To them, the value of any diskette is the same – full or empty – only a few dollars, or the value of the atoms . . . Our mind-set about value is driven by atoms. The General Agreement on Tariffs and Trade is about atoms. Even new movies and music are shipped as atoms. Companies declare their atoms on a balance sheet and depreciate them according to rigorous schedules. But their bits, often far more valuable, do not appear. Strange.35

This paradigm of trade and value – rooted in physical goods, or ‘atoms’– was at that very time on the point of a major pivot, the implications of which are still becoming apparent today. Technological change and the dramatic shift towards the free flow of information were putting pressure on the established copyright system, but were also removing information-based constraints on trade and creating entirely new ways of trading in knowledge products. Buyers and sellers could find each other much more readily, across the globe. New forms of widespread trading in intangible content were created, once it was possible when reaching substantial consumer markets to detach such content from the physical substrate or ‘atom-based’ proxy that had served as a vehicle for transactions in this content. The rapid dissemination of establishment of the Internet created the pathways for IP-protected content to be disseminated in the form of packets defined according to the TCP/IP protocol. It enabled the development of a unique, apparently seamless, multifunctional international trading environment, a digital counterpart to the borderless trading domain in the ‘analogue’ world that multilateral trade rules are crafted to support. A combination of this necessarily rigid, uniform protocol and a set of single, unambiguous addresses (the domain name system and the Internet Protocol (IP) logical addresses to which domain names point), creates a reliable system of exchanging and forwarding data to predictable addressees with confidence that the data will be intelligible to the recipient. The universality, the flexibility, the adaptability and openness

35

Nicholas Negroponte, ‘Bits and Atoms’, Wired, 3, 1 January 1995.

                                  



of the Internet – what Zittrain calls a generative grid36 – flows paradoxically from this rigid orthodoxy. The essence of the Internet is therefore a protocol, not any particular collection of hardware or network of data conduits – it is a rule rather than an assemblage of pluming. This provides an instructive metaphor for the more general function of international rules as underpinning greater flows of information, cultural exchanges and trade: Electronic commerce presents major policy contradictions for many governments. On one hand, the creation of the [TCP/IP] as a harmonised protocol for data exchange sets the stage for explosive growth of Internetbased commerce in a virtual global community and marketplace. On the other hand, the trade policies and strategies of governments are embedded in notions of territoriality . . . Governments must avoid fragmentation of the emerging global online market because of incompatible approaches to rule making among nations. The best way to do this is through negotiating effective multilateral rules. In key respects, the TCP/IP is a metaphor for good rule making. Just as the TCP/IP is a set of rules which makes globalised data exchange possible, so good trade rules facilitate global commerce, including electronic commerce.37

For the evolving trade in knowledge, the TCP/IP protocol was the equivalent, for intangible content defined in terms of its informational content, of the creation of the standard container for trade in physical goods. Thus the Economist has observed that, as a harmonised network for international exchanges, the Internet operates in a manner similar to an open international trade regime: ‘the internet is as much a trade pact as an invention . . . Just as a free-trade agreement between countries increases the size of the market and boosts gains from trade, so the Internet led to greater gains from the exchange of data and allowed innovation to flourish.’38 The rise of Internet commerce has illustrated how rules operate to promote international exchanges and, paradoxically, how a backbone of rules-based harmonization may provide a basis for diversity in cultural expression and diffusion of information. The trade policy significance was evident as early as 1998, when the WTO’s Geneva Ministerial Declaration on Global Electronic Commerce39 recognized the ‘new opportunities for trade’ that were then already unfolding. 36 37

38 39

Jonathan Zittrain, ‘The Generative Internet’ (1974), Harvard Law Review [2006]: 119. Department of Foreign Affairs and Trade, Driving Forces on the New Silk Road, Canberra: Department of Foreign Affairs and Trade, 1999, p. 85. ‘The Web’s New Walls’, The Economist, 2 September 2010. WT/MIN(98)/DEC/2, 25 May 1998.



 

The ‘atoms to bits’ shift predicted by Negroponte in 1995 has since eventuated to a considerable extent, transforming both domestic markets and international trade – not merely opening up new avenues for trading and thus competing in the supply of established goods and services, but also creating altogether new forms of commercial transaction, novel possibilities for exchanging value through new trading platforms. This transformation was most immediately apparent in the development of new practical means of supplying intangible content, notably the now widely-used platforms for consumer access to e-books and other publications, software applications, music and audiovisual works. But at a conceptual level, it redefined or at least diversified the very character of the commercial transactions that constitute trade in valuable intangible content, or the ‘purchase’ of such content (which is, in fact, generally the acquisition of a restricted use licence, defined by IP, by technological measures, by contract, or more typically by a combination of these). As debate continued about whether ‘digital products’ should be considered goods or services,40 or something else41, other business models emerged, blurring the boundaries of established forms of transaction in such IPdefined and IP-protected content, such as streaming subscriptions, ‘freemium’ shareware and in-app purchases. While many have the essential character of services, uncertainty remains over digital products that are analogous to traditional physical products containing essentially the same content. Equally, IP rights as such are increasingly traded as discrete items,42 whether through dedicated IP exchanges established as well-defined marketplaces,43 or in a broader sense of a ‘market’ for IP rights, such as the emerging ‘market for brands’, comprising purchase, franchising or licensing of trademarks and brands as the subject of 40

41

42

43

‘[T]here was still a lack of clarity with regard to the classification under GATT or GATS of certain products which can be delivered both in electronic form and on a physical carrier’. Summary by the Secretariat of the Issues Raised, Dedicated Discussion on Electronic Commerce under the Auspices of the General Council on 15 June 2001, WTO document WT/GC/W/436 (6 July 2001); despite continuing debate and a voluminous scholarly literature, the question remains unresolved at a formal level. See, e.g., WTO, Communication from Indonesia and Singapore, Preparations for the 1999 Ministerial Conference: Work Programme on Electronic Commerce, WT/GC/W/ 247, 9 July 1999. OECD, ‘Intellectual Property Market’, in OECD, Science, Technology and Industry Outlook, 2012, OECD Publishing, at dx.doi.org/10.1787/sti_outlook-2012-24-en. See, e.g., www.ip-marketplace.org/, maintained by the Danish Patent and Trademark Office; and Asia IP Exchange (AsiaIPEX) developed by the Hong Kong Trade Development Council, available at www.asiaipex.com.

                                  



commercial transactions in their own right,44 and the growth of markets for technology as such, ‘disembodied from physical goods’.45

Creating New Markets for IP The transformative effect of digital technologies on trade in knowledge products has been profound throughout the period of implementation of the TRIPS Agreement. The scale of this trade in IP is remarkable, as is its impact in many sectors. For many consumers, the most evident manifestation is in the form of online platforms for access to digital content, such as music downloads, apps, and e‑books. In 2001, the global recorded music industry46 reported revenues of USD 23.2 billion for sales of physical media; this total has fallen steadily since then, reaching USD 4.7 billion in 2018. Digital downloads only registered a significant proportion from 2004, peaked in 2012 at USD 4.3 billion, and fell to USD 2.3 billion in 2018. Streaming revenues rose sharply more recently, at USD 8.9 billion in 2018 well exceeding the revenues from both physical media and downloads. In the book sector, the Association of American Publishers reports that in the US market in 2018, for the second successive year, ‘publisher sales to online retail channels exceeded sales to physical retail channels with sales to online retail at $8.03 billion and sales to physical retail at $6.90 billion’.47 In a Chinese market of over USD 1 billion in 2016, digital editions reportedly accounted for around 2848 per cent. A survey of the publishing market in 2018 indicated the highest revenue share from digital editions in Japan (24.5 per cent), Sweden (23.2 per cent), the Republic of Korea (22.5 per cent) and the US (19.4 per cent).49 Despite only coming into existence with the introduction of the iPhone in 2008, app stores have become major commercial platforms 44

45

46

47 48

49

Carl Benedikt Frey, Atif Ansar and Sacha Wunsch-Vincent, ‘Defining and Measuring the “Market for Brands”: Are Emerging Economies Catching Up?’, Economic Research Working Paper No. 21, 2014. Ashish Arora and Alfonso Gambardella, ‘Ideas for Rent: An Overview of Markets for Technology’ (2010), Industrial and Corporate Change 19, 3: 775–803, available at https:// ssrn.com/abstract=1617051. International Federation of the Phonographic Industry (IFPI), Global Music Report 2019 // State of the Industry, available at www.ifpi.org/recording-industry-in-numbers.php. AAP StatShot Annual Report for Calendar Year 2018, 21 June 2019. WIPO, ‘The Global Publishing Industry in 2016: A Pilot Survey’, available at www.wipo_ ipa_pilotsurvey_2016. WIPO, The Global Publishing Industry in 2018, Geneva: World Intellectual Property Organization, 2020.



 

creating new international marketplaces for intangible content. One industry report reported that some 115 billion transactions took place in 2019 on the two major app stores, such as Google Play and the Apple iOS App Store.50 By one estimate, app store revenues were USD 120 billion in 2019, rising 110 per cent since 2016.51 This major area of international commercial activity is striking in its scale, given that this category of commerce did not exist before 2008; and yet much of this trade is not reported in existing trade statistics,52 the chief source of data currently the private sector itself, many details (such as the distribution of revenue to authors and app developers worldwide) being treated as confidential business information. A WTO report comments: It is not clear to what extent these transactions are recorded in current trade statistics, but their value is now a major component of revenues in the content industries, and a share of these earnings is redistributed to app developers, musicians, authors and other creators internationally. A clearer picture of these sizeable revenue flows would improve our understanding of the pattern of international trade in these sectors, and could lead to a more accurate understanding of how economies benefit from this form of international trade, as internet platforms serve to connect content developers across the globe with consumers in multiple jurisdictions.53

The WTO report examines the transformational effect of trade in digitizable goods (defined as physical goods that can be digitalized) including ‘cinematograph film; traditionally printed matter such as books, pamphlets, maps, newspapers, journals, periodicals, postcards and personal greeting message or announcement cards; video games; computer software; and recorded media such as musical records, tapes and other sound or similar recordings’54 and concludes that imports of such goods in physical form has fallen to 0.8 per cent of WTO members’ total imports, by contrast with the proportion of 2.9 per cent in 2000 (see Figure 2.3). 50

51

52

53

54

Sensor Tower, Q4 2019 Store Intelligence Data Digest, available at https://go.sensortower .com/Q4–2019-Data-Digest.html. App Annie, State of Mobile 2020, available at https://www.appannie.com/en/go/state-ofmobile-2020/. On the difficulties of measuring this trade, see Joscelyn Magdeleine and Andreas Maurer, ‘Measuring International Intellectual Property Transactions in a Globalized World: Current Challenges and Possible Improvements’, Chapter 5 in this volume, and Erick S. Oh, ‘The Global Digital Content Landscape’, Chapter 10 in this volume. WTO, World Trade Report 2018: The Future of World Trade: How Digital Technologies are Transforming Global Commerce, Geneva: WTO, 2018, p. 98. Ibid., p. 92.

                                  



Figure 2.3 Value and share of trade in physical form of digitizable goods Source: WTO, 2018.

The advent of widespread access to digital platforms has, indeed, created entirely new markets for content. One example is the market for caller tunes, the music played to those calling a mobile telephone. This is a major source of revenue for the predominantly domestic music sector in developing countries such as India, where, reportedly, over a billion transactions a day are undertaken domestically,55 and with revenue from foreign markets reportedly comprising 10 per cent of this income stream. Plainly, TRIPS negotiators could not have anticipated not merely the scale and importance, but even the very existence of this particular marketplace for highly specific licences for musical works. In contrast, with trade in such content when included on physical media, ‘purchase’ of digital content generally does not lead to a transfer of ownership, but rather simply entails a conditional and limited licence for access. Thus, the 2018 terms and conditions for Google Play refer to ‘purchase of content’ and a ‘sale contract’, but then stipulate: License to Use Content. After completing a transaction or paying the applicable fees for Content, you will have the non-exclusive right, solely as expressly permitted in these Terms and associated policies, to store, access, view, use, and display copies of the applicable Content on your 55

See www.comviva.com/news-events/comvivas-crbt-platform-delivers-1-billion-caller-tunesevery-day-india.htm.



  Devices or as otherwise authorized as part of the Service for your personal, non-commercial use only. All rights, title and interest in Google Play and Content not expressly granted to you in the Terms are reserved. Your use of apps and games may be governed by the additional terms and conditions of the end user license agreement between you and the Provider. Violation of License Terms. If you violate any of the Terms, your rights under this license will immediately terminate, and Google may terminate your access to Google Play, the Content or your Google Account without refund to you.56

Similarly, Apple’s Licensed Application End User License Agreement opens with the clear statement that ‘[apps] made available through the App Store are licensed, not sold, to you . . . Licensor grants to you a nontransferable license to use the Licensed Application on any Applebranded products that you own or control and as permitted by the Usage Rules. . .’57 And Amazon’s Kindle Store stipulates that ‘Kindle Content is licensed, not sold, to you by the Content Provider’, and grants ‘a nonexclusive right to view, use, and display . . . Kindle Content’ solely as permitted. A much broader policy debate surrounds this instance of the shift ‘from asset to access’, or the development of ‘an economy built around access relations’ shifting from ‘a property regime based on the idea of broadly distributed ownership to an access regime based on securing short-term limited use of assets controlled by networks of suppliers’.58 A recent critical analysis of the implications of trading ‘property rights for conditional privileges’ argues that the ‘baseline for property rights should be a function of the law, not contingent on the kindness of copyright holders and retailers’59 and calls for the law to recognize purchasers’ property rights over digital property, akin to property rights over chattels. For immediate analysis of the trade policy and economic implications, however, it remains the case that the very legal character of trade in digital content is, on the whole, significantly different from trade in corresponding physical carrier media, in terms of the formal transfer of ownership, in terms of the contingent character of access to content, 56

57 58 59

Google Play Terms of Service, 5 February 2018, available at https://play.google.com/ about/play-terms/index.html. See www.apple.com/legal/internet-services/itunes/dev/stdeula/. Jeremy Rifkin, The Age of Access, New York: Penguin Putnam, 2000, p. 4. Aaron Perzanowski and Jason Schultz, The End of Ownership: Personal Property in the Digital Economy, Cambridge, MA: MIT Press, 2016, p. 187.

                                  



and in terms of the further reach of applicable IP rights over the downstream use of and trade in the content to which access is procured.

. . . And New Trade Opportunities? At least at the level of principle, the resultant technological transformation of commercial practice has the potential both to improve prospects for equality of commercial conditions for enterprises around the globe regardless of their geographical location (the kind of market access that would be favoured, in principle, by trade policy), and to ensure more efficient, and more transparent, competition in domestic markets.60 The resultant new forms of trading, in principle, open up new prospects for socially beneficial competition and for international market access as a de facto form of trade liberalization. In particular, the interconnectivity afforded by the Internet protocol suite, has opened up new opportunities for participation in international commercial activity, notably by enterprises in developing countries and by micro, small and medium-sized enterprises (MSMEs). A recent position paper submitted by several WTO member governments observes that: Digital technology is transforming the global economy. This transformation presents new opportunities to promote inclusive economic growth, including by connecting rural to urban economies; opening new channels of trade for landlocked countries; facilitating the participation of women and micro enterprises in the formal economy; providing micro, small and medium-sized enterprises (MSMEs) access to a global consumer base; and facilitating cross-border trade in services previously considered not technically feasible.61

The statistics, such as they are, provide some evidence that developers across the world may be taking advantage of these opportunities. Apple claimed in January 2020 that its App Store ‘provides developers of all sizes access to customers in 155 countries. Since the App Store launched in 2008, developers have earned over $155 billion, with a quarter of those

60

61

D. Van Welsum, W. Overmeer and B. Van Ark, ‘Unlocking the ICT Growth Potential in Europe: Enabling People and Businesses’, The Conference Board for the European Commission, 2013. WTO, Trade Policy, The WTO and the Digital Economy, Communication from Canada, Chile, Colombia, Côte d’Ivoire, the European Union, the Republic of Korea, Mexico, Montenegro, Paraguay, Singapore and Turkey JOB/GC/116, JOB/CTG/4, JOB/SERV/ 248, JOB/IP/21, JOB/DEV/42 (13 January 2017).



 

earnings coming from the past year alone’.62 Alphabet reports that ‘over $80 billion has been earned by developers around the world from Google Play, [which has] over 2 billion active monthly users’.63 The practical implications for the creative industries – and development prospects – in a developing country context were set out in a recent ITC study considering the prospects for this sector in Rwanda,64 which concluded: digitization has contributed to the robust growth of creative industries in recent years, generating $2.25 trillion in revenue and 29.5 million jobs globally. It has also made digital export in the creative industries more accessible to small and medium-sized enterprises, including those from developing countries. Digital export in creative industries could underpin overall developing countries exports, providing a channel to leverage their rich culture and heritage for economic growth and diversification. Developing countries have already gained a foothold in the global creative export market. They account for 53% of worldwide exports of creative goods, amounting to $265 billion in export revenue. Although developed economies still dominate global trade in creative services, momentum is building in developing countries, with least developed countries’ share in exports of personal, cultural and recreational services growing 22.7% a year since 2012.

New platforms, tailored to specific needs, can open up these prospects further, for instance sites promoting local music such as Famemix in Rwanda, or Jokotext and MusikBi in Senegal. Despite these promising trends, considerable hurdles of course impede opportunities to benefit from these new avenues for trade in knowledge, particularly in the developing world, including shortcomings in digital infrastructure, lack of equitable access to established trading platforms, and the inevitable lag in the legal and regulatory framework in responding effectively to the distinctive characteristics of these new forms of trade. The following section reviews these evolving characteristics of intangible trade in knowledge products.

62

63

64

‘Apple rings in new era of Services following landmark year’, 8 January 2020, available at www.apple.com/newsroom/2020/01/apple-rings-in-new-era-of-services-following-land mark-year/. ‘Alphabet Q4 2019 Earnings Call, February 3, 2020’, available at https://abc.xyz/investor/ static/pdf/2019_Q4_Earnings_Transcript.pdf. International Trade Centre, Creative Industries in Rwanda: Digital Paths to Global Markets, Geneva: ITC, 2019.

                                  



But What Constitutes ‘Trade in BITS’? Regulating Trade-Related Aspects of IP in the Digital Environment The conclusion of a multilateral trade agreement on the ‘trade-related aspects’ of IP rights, in the form of the TRIPS Agreement, was, in effect, a recognition of the need to address what was already a complex interplay between international trade and the IP system; even at a time when that very interplay was in the process of radical transformation. Earlier, the IP system had been seen, conventionally, to be at odds with the objectives of market liberalization for trade in goods, so that IP protection had been conceived as an allowable exception under the GATT (Article XX), and not a positive obligation. When the TRIPS Agreement inverted this logic, and made application of high standards of IP protection a requirement within the WTO trade law system, critics of TRIPS viewed it as the intrusion of non-trade issues into the trading system in tension with the goal of market liberalization,65 by imposing ‘behind the border’ standards rather than opening up trade. Yet the preamble to TRIPS set the IP system in the context of international trade, recording WTO Members’ desire ‘to reduce distortions and impediments to international trade, and taking into account the need to promote effective and adequate protection of intellectual property rights, and to ensure that measures and procedures to enforce intellectual property rights do not themselves become barriers to legitimate trade’. Hence, the advent and implementation of the TRIPS Agreement spurred a reconsideration of the complex and diversifying interaction between IP and trade. Yet, emerging as it did from within the trade‑ingoods paradigm, the Agreement as originally conceived was not negotiated in anticipation of the massive, widespread tradability of IP rights as such, and did not envisage the sale of an IP licence as constituting the essential value at the heart of commercial transactions constituting international trade. The evolving diversification of commercial transactions dealing with intangible content, fuelled by the disruptive impact of digital technology – and the transformation both of content industries and of the opportunities for creative and innovative firms across the globe – suggests that it is timely to reconsider the ‘trade-related aspects’ of the IP system in the light of the dramatic transformations which have proceeded in parallel 65

E.g. A. Panagariya, ‘TRIPS and the WTO: An Uneasy Marriage’, in K. Maskus (ed.), The WTO, Intellectual Property Rights and the Knowledge Economy, Cheltenham: Edward Elgar Publishing, 2004, pp. 42–53.



 

with the implementation of the TRIPS Agreement across the WTO membership over the past quarter-century. These developments suggest that the evolution of the IP system and the digital environment together compel a reconsideration of what constitutes ‘trade’. Even as the ink had barely dried on the TRIPS Agreement, digital disruption of trade in IP content posed a range of challenges for the international IP system and above all for an agreement that sought to deal with the interplay between IP and international commerce – the pace of technological innovation effectively forcing a redefinition, in practice, of the ‘trade-related aspects’ of IP rights. The challenges posed confronted some fundamental concepts – on the face of it, at least, a kind of technological dilution or even abolition of national borders, with conceptual and practical difficulties in correlating the exercise and enforcement of IP rights with distinct and well-defined national jurisdictions, testing the longstanding principle of territoriality of IP rights and licences granted under them. Furthermore, the detachment of IP content from the physical carrier media that had conventionally served as a proxy for trade in IP rights made it possible for the general consumer to enter the market for pure licences in IP.

Trade in Digital Products Governments have long debated how WTO rules should apply to digital products: how they are taxed, valued and classified, and whether differential treatment of digital trade amounts to discrimination. For instance, what happens when the ‘same’ content is conveyed across the same border, but through electronic transmissions as a digital product, rather than as content embedded on a physically traded good? What is functionally the very same software package, for instance, may be downloaded or purchased on a disc; and a package purchased on a disc may be partially or fully upgraded over the Internet as part of the same commercial transaction. The packet‑switching technology and dispersed architecture of the Internet could mean that different elements of this content could pass undetected through many different jurisdictions. No distinct package or assemblage of atoms is presented to customs officials for clearance at a defined border. This uncertainty over the ‘real world’ characteristics of such products leads to the specific question of how digital products should fall within the conventional categories of goods and services. Early in the WTO debate, one delegation pointed out that ‘it was difficult to see how a distinction between “goods” and “services”

                                  



could be handled in practice, even if agreed on in theory. As the transmitted bytes of data streams consisted only of ones and zeroes, the delegation raised the questions how it was possible to decide for each individual case whether a particular transmission was covered by goods or services disciplines?’66 And, indeed, an early paper by Indonesia and Singapore, discussing the need for legal certainty in classification, speculated about the possibility of a third category of trade, that of trade in IP as such, which would be consonant with the increasing attention given to the contents, rather than the platform on which the contents are delivered: In attempting to classify digitized products one possible criterion that has been raised is to consider whether the product has a tangible counterpart in the physical world. This criterion could then be applied to over-thecounter purchases of books, music and software even if such purchases were delivered as digitized products and not in terms of their physical counterparts. An alternative is to just consider the contents themselves. Books, music and software are not in themselves new commercial products. It is just that prior to the advent of e-commerce, they were treated as goods because they had to be delivered in the form of a carrier media, be it paper, cassettes or diskettes etc. and those carrier media were classified as goods. Now that those forms of tangible carrier mediums are no longer necessary maybe what we need to consider is whether the software and music would continue to be classified as goods, or it might be more appropriate for them to be classified as services. It may also not be a coincidence that all these three examples, without a carrier medium are intangible goods considered under the ambit of intellectual property rights. Could such products then be simply considered as trade in intellectual property rights and not be classified as a good or a service? What is paramount though is that the criteria for classification should provide legal certainty on how the good, the service or the intellectual property right is to be treated.67

Meanwhile, the trade in digital products has grown apace and a massive body of practical experience has accumulated. This may indicate either that no definitive position is needed, is feasible or desirable under international trade law; or it may imply that the question may ultimately be resolved in a more ad hoc way, through pragmatic negotiation, through dispute settlement or through the accumulation of bilateral 66

67

General Council, ‘Interim Review of Progress in the Implementation of the Work Programme on Electronic Commerce, Communication from the Chairman of the Council for Trade in Goods’, WT/GC/24, 12 April 1999, p. 3. WTO document WT/GC/W/247, note 41 above, paras 10–14.



 

and regional trade agreements dealing with digital products, with potentially diverse forms of classifications applicable to a widening array of such products. To be sure, WTO dispute settlement has effectively confirmed that services delivered over the Internet are indeed covered by GATS obligations, despite the positions taken by some governments in WTO policy discussions that either an express common understanding would be desirable, or specific, additional commitments concerning Internet-supplied services would be required before any obligations came into effect.68 Recent bilateral and regional trade agreements have, in any case, addressed the issue in different ways.69 A number of recent agreements have established rules expressly addressing trade in digital products, defining such products and regulating their trade in terms that are similar, but not identical. In general, such agreements indicate, in various formulations, that any definition of a digital product is without prejudice to Parties’ views on whether trade in digital products through electronic transmission should be categorized as trade in services or goods. But the scope of definition of ‘digital products’ differs, including on apparent coverage of content included on carrier media. The Comprehensive and Progressive Agreement for Trans-Pacific Partnership (CPTPP)70 and the recent agreement between Canada, Mexico and the United States (variously styled by its parties T-MEC, CUSMA and USMCA)71 both define a digital product as ‘a computer programme, text, video, image, sound recording or other product that is digitally encoded, produced for commercial sale or distribution, and that can be transmitted electronically’. The Australia–Japan Economic Partnership Agreement defines the term similarly but expressly excludes products ‘that are fixed on a carrier medium’72 (in effect, the ‘chattels’ discussed above). By contrast, the Republic of Korea–US Free Trade Agreement defines digital products in otherwise identical terms, but ‘regardless of whether they are fixed on a carrier medium or transmitted electronically’.73 Given that the principal 68

69

70 71 72 73

See the summary of the debate in Sasha Wunsch-Vincent, ‘The Internet, Cross-Border Trade in Services, and The GATS – Lessons from US-Gambling’, World Trade Review 5, 3: 319–356, p. 323. For a general discussion of trade agreements in this area, see Mira Burri, ‘Adapting Trade Rules for the Age of Big Data’, Chapter 20 in this volume. Article 14.1. Article 19.1. Article 13.2. Article 15.9.

                                  



undertaking applied to trade in digital products is one of nondiscrimination, such divergences in definitional scope may have limited practical impact – beyond reinforcing a general liberalization of such trade – but they clearly betoken some divergence as to the trade law significance of digital content embedded on a traded physical carrier medium. Since, also, some provisions require treatment of digital products that is no less favourable to ‘like’ digital products, to the extent that the definition includes content on carrier medium, this nondiscriminatory principle may have implications for traded ‘chattels’ that have traditionally been treated as traded goods – particularly if physical carrier media are subject to tariffs or other constraints at the border which are not applied to their intangible counterparts. The Comprehensive and Economic Trade Agreement (CETA) between the EU and Canada takes a different approach to dealing with trade in parcels of digital content. It does not refer to digital products, but, in its e-commerce chapter, defines a ‘delivery’ as ‘a computer program, text, video, image, sound recording or other delivery that is digitally encoded’, a phrasing closely analogous to that of other agreements’ definition of ‘digital product’. Despite the substantive overlap in these definitions, this choice of terminology seemingly emphasizes the means of transmission, and not the treatment or categorization of digital content as such. Indeed, the term ‘delivery’ may perhaps allude to a trade in services framework.74 In any event, as in other trade agreements, a key element concerns the exclusion of ‘a customs duty, fee, or charge on a delivery transmitted by electronic means’. Chapters on electronic commerce or digital trade in recent trade agreements are generally framed with the proviso that their terms are without prejudice to the parallel rules on IP. If the protection of IP – and the approach taken to the question of exhaustion – is to prevail in this way over parallel, newly crafted obligations to treat digital products (including, potentially, ‘like’ digital products traded on physical carriers), this may provide a basis for differential treatment between intangible and tangible versions of the same content in some circumstances. More generally, and at the very least, this kind of provision may betoken the need for a more coherent view, from a trade policy point of view, of the IP dimension of transactions for digital products. 74

For instance, Article XXVIII of the WTO General Agreement on Trade in Services defines supply of a service to include its ‘delivery’: (c) ‘measures by Members affecting trade in service include measures in respect of . . .’.



 

Beyond Trade in Goods? Taken together, these developments – in technology, in business patterns (particularly new ways of linking content providers with content consumers) and in international agreements – raise probing questions about the very character of trade, its essential legal and economic characteristics, and how it is to be measured and classified. In particular, the development of international digital platforms that permit the trading of pure content means that the commercial nature of the transaction is decreasingly less likely to be defined by the transfer of ownership in a physical thing; the true character of the transaction emerges more clearly as the acquisition of a licence or contract over content defined by IP rights. The WTO’s multilateral trade agreements – and the overarching Agreement Establishing the WTO itself – do not venture an express formal definition of what constitutes ‘trade’ as such. A dispute settlement panel has observed that ‘although the preamble to the WTO Agreement refers in particular to expanding trade in goods, it does not give sufficiently precise information on the terms which we must interpret’.75 And this may be inevitable, as trade policy has long wrestled with the boundaries of the kind of commercial transactions that are validated as ‘trade’, which in turn raises deep questions of economic value. By one recent, standard definition, ‘trade’ is the ‘exchange of goods between two individuals or nations . . . the basic component of economic activity . . . undertaken for mutual advantage’.76 Naturally, such a definition begs the question of what is counted as a ‘good’ – and how material and enduring it must be. The intuitive tendency to associate economic value with the outcomes of productivity only when captured in lasting tangible form – the proverbial ‘things you can drop on your foot’ – is already apparent in Adam Smith’s reference, in the Wealth of Nations, to the intangible or ephemeral product of ‘players, opera-singers, operadancers, etc’. as producing ‘nothing which could afterwards purchase or procure an equal quantity of labour. Like the declamation of the actor, the harangue of the orator, or the tune of the musician, the work of all of them perishes in the very instant of its production’.77

75 76 77

Panel Report, EC – Asbestos, para. 8.48. Oxford Dictionary of Economics, 4th ed., Oxford: Oxford University Press, 2013. Adam Smith, An Inquiry into the Nature and Causes of the Wealth of Nations, Henry Frowde (ed.), Oxford: Oxford University Press, 1909 (1776).

                                  



In more recent times, the increasing recognition of the value of the intangible component of commercial transactions – and its consequential significance for economic growth and trade policy – inevitably led to a reframing of what constitutes ‘trade’, resulting most consequentially in the incorporation of trade in services within the framework of multilateral trade law through the vector of the General Agreement on Trade in Services,78 and parallel developments in numerous bilateral and regional trade agreements. Thus, in Goode’s dictionary, trade ‘usually refers to the sale and distribution of goods and services across international borders. There are many different ways of doing this, but there must be a commercial element for a transaction to qualify as trade’. By one definition, a good is ‘any physical object, natural or manmade, or service rendered, that could command a price in a market’.79 A more precise technical definition of tradeable goods in the context of statistics is found in the recommendation that ‘international merchandise trade statistics record all goods which add to or subtract from the stock of material resources of a country by entering (imports) or leaving (exports) its economic territory’.80 The definitions of these concepts begin to count – literally count – when used to define statistics. In balance-of-payments statistics, goods are defined as ‘physical, produced items over which ownership rights can be established and whose economic ownership can be passed from one institutional unit to another by engaging in transactions . . . The production of a good can be separated from its subsequent sale or resale’.81 And services are defined as ‘the result of a production activity that changes the conditions of the consuming units, or facilitates the exchange of products or financial assets. Services are not generally separate items over which ownership rights can be established and cannot generally be separated from their production’.82 Notably, there are certain cases in which the passage of physical goods across borders is counted as a service, not because of their inherent nature or value, but because of the nature of the commercial transaction that leads to their shipment, specifically ‘newspapers and periodicals sent 78

79 80

81

82

See Lee Tuthill, Antonia Carzaniga and Martin Roy, ‘How Digitization is Transforming Trade’, Chapter 3 in this volume, for a general overview of the trade in services dimension. G. Bannock, R. E. Baxter and E. Davis, Dictionary of Economics, London: Penguin, 2004. United Nations, International Merchandise Trade Statistics: Concepts and Definitions 2010, ST/ESA/STAT/SER.M/52/Rev.3 Statistical Papers Series M No. 52, 12. IMF, Balance of Payments and International Investment Position Manual, Washington, DC: International Monetary Fund, 2009, p. 119. Ibid.



 

under direct subscription’ and ‘media used for carrying software customized or written for a specific client or originals of any nature, where identified’.83 In other words, whether a transaction is classed as a good or service in these cases will in turn depend not on the characteristics of the physical good itself, but rather the commercial context within which the good is shipped. This is a telling reminder that, again, it is the character and legal context of the way in which value is defined and exchanged that determines the nature of ‘trade’, rather than the objective characteristics of any object, tangible or otherwise, that is transmitted in the course of, or as a consequence of, the transaction. The extent to which IP rights define and govern transactions in both physical and digital products can significantly shape both domestic and international trade in these materials, specifically when physical goods embody IP-protected content and when digital products are defined and transacted in the form of IP licences. Hence the applicable IP rights can form an inherent part of the context of commercial transactions and in turn define the very nature of the market for such products. Their significance is particularly evident when considering the effect of exhaustion of IP rights, and more generally as to whether conceptual and legal preference is to be given to the tangible character of traded goods, or whether the IP dimension is to prevail. On the domestic plane, the interplay between IP and trade in goods was considered by the US Court of Appeals for the Federal Circuit in the ClearCorrect case84, concerning the authority of the International Trade Commission under the Tariff Act to ‘remedy only those unfair acts that involve the importation of “articles” as described in 19 USC § 1337(a)’. The case concerned whether this authority extended to the electronic transmission of digital data, when a digital data model for the alignment of teeth, transmitted from Pakistan to the United States, was argued to be an importation of an article infringing a US patent and thus subject to border measures as a form of unfair competition. The ITC had held that ‘the digital data sets at issue . . . are true articles of international commerce that are imported into the United States, and their inclusion within the purview of section 337 [of the Tariff Act] would effectuate the central purpose of the statute’.85 However, on appeal to the Federal 83 84

85

International Merchandise Trade Statistics: Concepts and Definitions 2010, note 80, 21. ClearCorrect Operating, LLC v. International Trade Commission, 810 F.3d 1283 (Fed. Cir. 2015). Ibid., p. 1304.

                                  



Circuit, the majority opinion found that ‘articles’ means ‘material things’, acknowledging that ‘electronic transmissions have some physical properties—for example an electron’s invariant mass is a known quantity’86 – but arguing that ‘commonsense dictates that there is a fundamental difference between electronic transmissions and ‘material things’. The majority saw this as a policy matter for Congress, which ‘is in a far better position to draw the lines that must be drawn if the product of intellectual processes rather than manufacturing processes are to be included within the statute’.87 In her dissent, Judge Newman argued that the Tariff Act was ‘enacted to provide additional support to domestic industries that dealt in new and creative commerce, by providing an efficient safeguard against unfair competition by imports that infringe United States patents or copyrights’.88 She argued that the digital datasets were patentable inventions: ‘[i]t is now beyond debate that digital goods are subject to the patent law, and it is beyond debate that digital goods can be imported’ and ‘the intention to omit unforeseen, later-discovered technologies cannot be imputed to this statute’.89 This case highlights one of the central trade policy questions in considering the implications of digital disruption: whether an intangible product, with similar properties for the consumer as its tangible counterpart, should be treated in the same way, or should be treated differently, on the basis that it is essentially a bundle of intangible rights, with no physical substrate that can be ‘owned’ as a chattel; or, more succinctly, how far we can indeed apply the majority, avowedly ‘commonsense’, view in ClearCorrect that there is indeed ‘a fundamental difference between electronic transmissions and material things’ (emphasis added).

IP in the Digital Marketplace: The Effect of Exhaustion One consequential question for trade and IP policy is whether, to what extent and how, digital disruption can and should create single markets for content that transcend national borders. Can digital markets for content defined by IP rights be forged internationally, for the mutual 86 87

88 89

Ibid., p. 1287. Ibid., p. 1302, citing its decision in Bayer AG v. Housey Pharm., Inc., 340 F.3d 1367, 1374 (Fed. Cir. 2003). Ibid., p. 1304. Ibid., p. 1307.



 

benefit of the creators and consumers of creative content, while still compatible with IP systems defined and structured on a territorial basis? Digital disruption has created practical possibilities for global markets for IP-protected digital content, and to some extent technological development has enabled the de facto eradication of the national borders that used to enable IP right holders to segment markets for IP content according to geographical boundaries. Indeed, this leads to competing expectations – right holders seeking to maintain downstream control over the distribution of content to other jurisdictions, consumers of content expecting immediate access to a global single market of content, with a corresponding single global licence. In the past analogue context, we have seen how the US Supreme Court drew an analogy between the purchaser’s rights over the imported copyright-protected textbook and ownership of a ‘chattel’ (a tangible item of property), rather than viewing the physical book as being essentially a mere carrier medium for protected content and licensed IP. And in earlier US jurisprudence, a copyright restriction on radio broadcast of a sound recording was construed as an ‘equitable servitude on a chattel’,90 seeing the essential nature of the transaction as ownership of the physical object, while ‘enjoyment of the possession of a physical copy of the recording was subject to continuing liability to the originator of the sound recording’.91 The advent of digital technologies, and the trading in digital copies of copyright works in particular, inevitably leads to a recalibration of the market and a redefinition of the relationship between content producer, downstream trader and consumer. How should the treaty negotiator, the domestic legislator, the court respond? Is this a technical matter of applying established principles in a novel context, or are new or adjusted principles needed? The approach taken to the exhaustion of rights over purely digital content provides an instructive test case: does the right holder have a different entitlement to reach through the original transaction and to control downstream distribution, depending on whether the content is carried on physical media or is despatched as bundles of data on a packet-switching network?

90 91

Waring v. WDAS Broad. Station Inc., 194 A. 631, 638 (1937). Antony Taubman, ‘TRIPS Encounters the Internet: An Analogue Treaty in a Digital Age, or The First Trade 2.0 Agreement?’, in Mira Burri and Thomas Cottier (eds), Trade Governance in the Digital Age, Cambridge: Cambridge University Press, 2015, p. 314.

                                  



On‑demand delivery of digital copies of copyright material was not directly addressed in the TRIPS negotiations, which had concluded effectively in 1991, given the acknowledged lack of awareness among negotiators as to the impending impact of the Internet on the content industries. Thus, how to regulate on-demand delivery of protected works, performances and sound recordings became a central issue in the work at WIPO, which resulted in the adoption of the two new copyright treaties in December 1996, the WCT and WPPT. The principal purpose of these so-called ‘Internet Treaties’ was to adapt international rules for the protection of copyright and the rights of performers and producers of sound recordings to the digital revolution, in particular, the distribution of copyright material over the Internet. WCT Article 8 on ‘Right of Communication to the Public’ is the most important element in the WCT building on the platform established by TRIPS. It is intended to cover on-demand delivery of protected works over the Internet, a scenario captured in the carefully negotiated wording ‘the making available to the public or their works in such a way that members of the public may access these works from a place and at a time individually chosen by them’. The Basic Proposal, commenting on what became Article 8, explains the different scope and effect of the distinct rights of communication and of distribution: It should be pointed out that no rights are exhausted in connection with communication to the public. Should communication of a work result in the reproduction of a copy at the recipient end, the work may not be communicated further to the public or distributed to the public without authorization. Exhaustion of rights is only associated with the distribution of tangible copies.92

The text of the provision as contained in the Basic Proposal was adopted without any changes, and a similar approach was followed in WPPT Articles 10 and 14 (which deal respectively with the analogous right of making available fixed performances and phonograms). The 2012 Beijing Treaty on Audiovisual Performances in its Article 10 (Right of Making Available of Fixed Performances) extended this right also to performers of performances fixed in audiovisual fixations. The overall effect is to create different conditions for the further sale of IP-protected content, depending on whether it is provided as a distribution or a communication.

92

Basic Proposal, CRNR/DC/4, para. 10.20.



 

At the level of domestic jurisprudence, the European Court of Justice (ECJ) has concluded that ‘exhaustion of right applies to the tangible object into which a protected work or its copy is incorporated if it has been placed onto the market with the copyright holder’s consent’.93 However, further commercial use – in this case, transfer to a new medium – is still caught by the copyright owner’s rights, as it amounts to a new form of commercial use not covered by the original copyright licence. A more recent ECJ decision considered the implications of the distinction between physical and digital transmission of copyright works. The Tom Kabinet case94 concerned a system of trading in e-books, which entailed selling second-hand e-books which had been legitimately purchased from the publishers. In addressing whether the purchaser of an ebook was entitled to resell it in this way, the case turned on whether, in the terms of Directive 2001/29/EC (the Infosoc Directive, which directly applied the WCT terminology) the supply of an e-book was a ‘distribution to the public’, akin to the second-hand sale of a physical book (implying exhaustion of rights associated with the further sale, since it was assumed that the right holder received sufficient remuneration from the first sale), or a ‘communication to the public’ (for which the WCT had expressly precluded exhaustion of rights) – in the latter case, the first sale of the e-book would not bring with it the entitlement of the purchaser then to sell the e-book to a third party. The ECJ concluded that ‘the supply to the public by downloading, for permanent use, of an e-book is covered by the concept of “communication to the public” and, more specifically, by that of ‘making available to the public of [authors’] works in such a way that members of the public may access them from a place and at a time individually chosen by them’. This ruled out the exhaustion of rights following the initial supply of the e-book. The Court’s decision was based on its interpretation of the Infosoc Directive, but was supported by policy considerations that emphasized the distinct economics of trade in hard copies and in e-books. Thus the court analysed both the WCT and the travaux préparatoires of the Infosoc Directive itself to conclude that the legislators had intended the rule of exhaustion to apply only to distribution of tangible objects incorporating the copyright work, such as printed books on a material medium, and not to the online communication of works: 93 94

All Posters, C-419/13, 22 January 2015, curia.europa.eu/juris/liste.jsf?num=C-419/13. European Court of Justice, Decision of 19 December 2019, C-263/18 – Nederlands Uitgeversverbond et al. v. Tom Kabinet Internet BV et al. (‘Tom Kabinet’).

                                  



interactive on-demand transmission was a new form of exploitation of intellectual property, in relation to which the Member States were of the view that it should be covered by the right to control communication to the public, while stating that it was generally accepted that the distribution right, which applies exclusively to the distribution of physical copies, does not cover such transmission.95

The court set out the policy distinction between digital and material supply of books in these terms: The supply of a book on a material medium and the supply of an e-book cannot, however, be considered equivalent from an economic and functional point of view . . . dematerialised digital copies, unlike books on a material medium, do not deteriorate with use, and used copies are therefore perfect substitutes for new copies. In addition, exchanging such copies requires neither additional effort nor additional cost, so that a parallel secondhand market would be likely to affect the interests of the copyright holders in obtaining appropriate reward for their works much more than the market for second-hand tangible objects, contrary to the objective [of establishing a high level of protection of authors, allowing them to obtain an appropriate reward for the use of their works, including when a communication to the public takes place].96

The result had been different in an earlier ECJ case addressing digital exhaustion, the UsedSoft case, which turned on whether digital copies of software, legitimately purchased as downloads (not on physical media such as discs), could be resold to third parties. In that case, the Court had concluded that the right of distribution could be exhausted when the software is sold either on a physical carrier medium (such as a CDROM) or as a software download with an unlimited period of use: the right of distribution of a copy of a computer program is exhausted if the copyright holder who has authorised, even free of charge, the downloading of that copy from the internet onto a data carrier has also conferred, in return for payment of a fee intended to enable him to obtain a remuneration corresponding to the economic value of the copy of the work of which he is the proprietor, a right to use that copy for an unlimited period.97

It therefore concluded that An author of software cannot oppose the resale of his ‘used’ licences allowing the use of his programs downloaded from the internet. The 95 96 97

Ibid., para. 43. Ibid., para. 58. UsedSoft, C-128/11, 3 July 2012, available at www.curia.europa.eu/juris/liste.jsf?num=C128/11.



  exclusive right of distribution of a copy of a computer program covered by such a licence is exhausted on its first sale.

This would only apply, however, if the nature of the transaction was effectively a true sale of the program, with transfer of ownership, even if it was formally structured as the acquisition of an IP licence. The transfer, thus, must be definitive, in that the transferor (or first proprietor) must delete the original copy, and the transfer must correspond with the same bundle of rights as initially transferred. The Court cites a definition of a ‘sale’ as ‘an agreement by which a person, in return for payment, transfers to another person his rights of ownership in an item of tangible or intangible property belonging to him’.98 The TomKabinet decision explains why the outcome is different for computer programs as against e-books. Computer programs are dealt with by a lex specialis, the Computer Programs Directive 2009/24, which expressly provided for the exhaustion of rights regardless of whether computer programs are supplied as downloads or in tangible form. It found that it was ‘abundantly clear the intention of the European Union legislature to assimilate, for the purposes of the protection laid down by [the Software Directive], tangible and intangible copies of computer programs.’ And, expounding the policy rationale for this distinction, it clarified that ‘from an economic point of view, the sale of a computer program on a material medium and the sale of a computer program by downloading from the Internet are similar, since the online transmission method is the functional equivalent of the supply of a material medium. Accordingly, interpreting Article 4(2) of Directive 2009/24 in the light of the principle of equal treatment justifies the two methods of transmission being treated in a similar manner’. From a policy perspective, therefore, the Court justifies the distinction in terms of the inherently different economic character of markets for books and for computer programs: books naturally deteriorate with further use and distribution, while computer programs do not. These considerations are assumed to be factored into the remuneration originally received by the right holder upon the first sale. Thus these contrasting decisions – applying different treatment to the extent to which two digital products can be further traded after purchase – hinge on interpretation of both treaty and statute, and on policy and economic rationale. Within the domestic market in the United States, it is well established that physical copies of recorded music or of software, for instance on 98

Ibid., p. 42.

                                  



optical discs, can be sold on to third parties: the Supreme Court confirmed this in the two decisions discussed above, with reference to the restraints on the alienation of chattels. Digital files can be resold if they are carried on a physical platform such as an MP3 player.99 But what if the IP content is not carried on a physical chattel? The ReDigi case100 dealt with the application of the US ‘first sale’ doctrine to a marketplace developed for the reselling of lawfully purchased and downloaded digital music files. ReDigi had developed a market for such files which operated by transferring files to the purchaser while deleting them from the seller’s computer, so that after ‘data migration’ the seller would no longer retain a copy of the digital file that they had sold. US copyright law (USC Section 109(a)) sets out the ‘first sale doctrine’ as follows: ‘the owner of a particular copy or phonorecord lawfully made under this title, or any person authorized by such owner, is entitled, without the authority of the copyright owner, to sell or otherwise dispose of the possession of that copy or phonorecord’. The case therefore turned in part on whether reselling digital music files fell within this exception to copyright – in effect, whether the right to object to such reselling had been exhausted (as it undoubtedly has for copyright works sold on physical media). The trial court and appeals court considered a range of legal and policy questions in reaching the conclusion that such reselling, against the authority of the original right holder, was not permitted under US copyright law. In essence, the finding was that the data migration process necessarily entails making copies of the digital files, copies that are not authorized by the right holder. When the purchaser downloads a file to their data storage, that storage becomes a new, unauthorized phonorecord. A range of policy issues were raised by the defendants, and in amici briefs, notably on the economic benefits that would flow from an effective secondary market for digital files. The appeal court declined to venture into issues which it saw as policy questions to be settled by Congress, but nonetheless pointed out that ‘the establishment of ReDigi’s resale marketplace would benefit some, especially purchasers of digital music, at the expense of others, especially rightsholders, who, in the sale of their merchandise, 99

100

US Copyright Office, ‘A Report of the Register of Copyrights Pursuant § 104 of the Digital Millennium Copyright Act ’, 2001, available at www.copyright.gov/reports/studies/ dmca/sec-104-report-vol-1.pdf. Capitol Records, LLC v. ReDigi Inc., No. 16-2321 (2nd Cir. 2018), available at https://law .justia.com/cases/federal/appellate-courts/ca2/16-2321/16-2321-2018-12-12.html.



 

would have to compete with resellers of the same merchandise in digital form, which, although second hand, would, unlike second hand books and records, be as good as new’.101 Further, the court argued that a secondary market could be sustained by selling digital files on storage media such as thumb drives. But concerning a market purely in digital content as such, the court concluded that ‘[i]f ReDigi and its champions have persuasive arguments in support of the change of law they advocate, it is Congress they should persuade. We reject the invitation to substitute our judgment for that of Congress’. The extent of ‘digital exhaustion’ – and the rationales for it to converge with or diverge from exhaustion of rights over content embedded in physical carrier media (‘chattels’) – are emblematic of the unresolved policy tensions for the IP system and international trade, in the light of the technology-driven trend towards a global marketplace for knowledge products, on the one hand, and the maintenance of territorially-based systems for defining and giving effect to IPRs, on the other. Even as the US courts have favoured international exhaustion of IPRs when embedded in physical carrier media, potentially opening up international trade for such knowledge products in the form of a secondary market, the question of whether purely digital content can be traded the same way also appears to be unsettled, with divergent views in two major jurisdictions, the EU and the US. And the ClearCorrect case has shown the difficulty of defining the perceived boundaries, even within the law of one jurisdiction, of what can be construed as a traded article, when it is no longer necessary to rely on a physical carrier medium as a platform for valuable content to be traded internationally.

Unbundling the Packets: Analysing the IP Dimension of Trade in Digital Products ‘Big Data’ or Small Packages? This chapter has explored how rights over valuable intangible content can shape the nature and extent of markets both for physical media and for digital products. This interlacing of IP law and the law governing trade in goods means that decisions on the formulation and application 101

Plaintiffs-Appellees, 15 16 v. 17 18 ReDigi Inc., John Ossenmacher, Larry Rudolph, AKA Lawrence S. Rogel, No. 16-2321, Document 188-1, p. 34, available at www.supremecourt .gov/DocketPDF/18/18-1430/90519/20190304145955525_00000003.pdf.

                                 



of IP law and policy taken by trade negotiators, by legislators and by the courts can significantly shape the structure of domestic and international trade. We cannot therefore understand the nature of trade and the contemporary international marketplace without analysing the IP component, and the extent to which a licence or contract under an IP right defines the scope of legitimate trade in physical and digital products. The most immediate observation that can be derived from the growth of trade composed essentially by IP licences is that it is erroneous to consider ‘trade in bits’ as simply a cross-border flow of packets of data analogous to the flow of containers of traded goods. Discussions about the trade policy implications of e-commerce have underscored the need to differentiate different forms of data flow. An examination of the characteristics of digital-enabled ‘trade in IP’, the trade in ‘digital products’ and digital ‘deliveries’ demonstrates the difficulties of defining, regulating, tracing and measuring such trade. Given the designedly dispersed character of the Internet as a widely distributed packet-switching network, most digital products are not sent as clear, bounded bodies of data in a single, discrete transmission over a dedicated, permanent connection between source and destination. The flow of data is complex, heterogenous in content and subject to distinct legal constraints. Consider the data packets involved in a single Internet transaction over the purchase of a single downloaded song, app or e-book, by a consumer in one jurisdiction from a commercial site in another, noncontiguous, jurisdiction. These packets of data, all defined by the TCP/IP protocol, would typically include: • data on available content for purchase from the site, and the applicable terms and conditions (information that is normally freely accessible to the public over the Internet, although not all of it necessarily in the public domain) • information concerning the purchaser, including personal details and credit card details (subject to privacy and confidentiality considerations) • transmission from the purchaser to the supplier, for instance legally confirming adherence to licensing terms and conditions and ostensibly entering into a contract • metadata concerning the digital product or downloaded file • rights management information relating to copyright content (information which itself may be protected in accordance with the requirements of the WCT and similar standards in bilateral agreements) and



 

• the data packets constituting the actual digital file for which access and a licence to use have been purchased (normally the subject of IP rights or contractual obligations constraining its field of use and subsequent dissemination, copying and possibly onward sale, depending on the applicable exhaustion regime). In a normal transaction – one of the uncountable billions of such transactions taking place each year – none of these data would pass from vendor to purchaser as a discrete bundle or package, or a single ‘file’. Indeed, the very design and logic of the Internet not only makes this unlikely, but makes it difficult to predict just what pathway across the Internet individual packets of data would take. The digital product itself – such as the mp3 file, e-book file, – is dispatched through numerous data packets over the Internet. These packets of digitized information may follow different routes to reach their destination, potentially passing through different territories, before being reconstructed within the digital device of the purchaser once the packets are all received. Thus the ‘digital product’ or ‘delivery’ referred to in recent trade agreements normally only exists in a coherent package in the memory of devices at either end of the transaction – plainly, none of the individual packets of data variously transmitted between the two end nodes will constitute transmission of the digital product or full delivery as such. Moreover, the transaction may entail a continuing right to download a purchased digital product perpetually, or for a distinct period or number of times. What is ultimately significant, therefore, in understanding this transaction is not the complex and diverse flow of data packets, but rather the relationship between the purchaser and the vendor, and the nature of the rights to use data that are transferred or licensed in this process and the limitations on those rights, and the transfer of ownership (if any), however ownership is construed. It is natural to speak of having ‘purchased’ an e-book, a song or an app, and some authors have argued that terms such as ‘buy’ or ‘purchase’, when applied to digital transactions, may be positively misleading.102 And it is counterintuitive to suggest that the essence of the commercial transaction – the ‘trade’ – involving a digital product does not constitute the electronic transmission of data as such, but rather the legal relationship between the purchase of a limited IP licence or contract. Yet even

102

See, for instance, Aaron Perzanowski and Jason Schultz, n. 59 supra, 84–101.

                                 



when ‘analogue’ transactions were based on the purchase of physical carrier media, it was generally fallacious to assume that transfer of ownership of the physical platform – the book, the disc, the tape – brought with it unhindered rights to use the embedded content: for instance, purchase, and ownership, of a CD or DVD did not confer a right of public performance of the musical or cinematic work it carried. And the purchase of content in foreign jurisdictions also did not bring with it an entitlement to trade it in jurisdictions that applied national exhaustion to the exercise of IP rights, even when the content was embedded in a physical carrier medium, a chattel. Thus it is suggested that it is impossible to reach a clear understanding of the inherent characteristics of many digital transactions without information on the legal relationship, including IP licences and contractual undertakings, that is established between the supplier and the purchaser. While statistics may be relatively accurate, if never fully complete, for business to business transactions, it becomes extremely difficult to measure or even gauge the full extent of business to consumer licences for IP that number in the billions internationally within a host of different and evolving business models.

Shedding Light: A Quantum Theoretical Approach to IP? The question of whether digital products should be classed as goods or services has not been definitively resolved at the multilateral level, whether by agreement or in the course of dispute settlement. The complex character of digital products, as outlined above, and the different categories of data despatched over networks, may give some illustration of the inherent difficulty in reaching a definitive position on this matter without considering the specific practical circumstances and legal context of digital products – just as the international despatch of some physical goods can yet be measured as part of services trade in some circumstances (as cited in GATT and the ‘trade in goods’ paradigm above), when the legal character of the underlying transaction makes this the more appropriate choice. And, indeed, as noted, a number of recent trade agreements have defined and regulated ‘digital products’ expressly without taking a position on whether they should be classed as goods or services (Trade in digital products, above). Add to this complex analytical framework the differing potential effects of applicable IP rights and the shortcomings of an overly deterministic approach become evident. An



 

improbable metaphor – that of the conception of light in quantum theory – may unexpectedly assist in creating a more enabling framework for understanding these diverse forms that IP may take in practice. A central insight of the quantum theory view of light is wave-particle duality – light can exhibit the behaviour of both waves and of particles or packets (‘quanta’), and may be observed and measured in either form. Concerning wave-particle duality, Einstein and Infeld observed: It seems as though we must use sometimes the one theory and sometimes the other, while at times we may use either. We are faced with a new kind of difficulty. We have two contradictory pictures of reality; separately neither of them fully explains the phenomena of light, but together they do.103

Without placing any significant emphasis on this admittedly tenuous metaphor, it suggests an analogous way of viewing IP rights, and indeed the diverse packets of data that are despatched in association with the delivery of digital products. Rather than assuming one single, exhaustive position as to how to categorize the essential character of IP rights and digital products in today’s technological and economic context and in a diversifying international trading system, it would be better to accept that they have a range of different properties in practice, which can be observed variously from different vantage points and will be more or less pertinent to the legal and commercial context of trade. Thus the grant, regulation and enforcement of the very same IP right may, in principle, be conceived in the international legal space as • a conditional exception to trade law disciplines, as a kind of behind‑the‑border domestic regulatory measure that may need to be justified under trade law as non-discriminatory and not traderestrictive 104 • a component of market access, and a means of avoiding distortions 105 to trade

103

104

105

Albert Einstein and Leopold Infeld, The Evolution of Physics, Cambridge: Cambridge University Press, 1938, p. 278, available at https://archive.org/details/evolutionofphy si033254mbp/. Economic and Trade Agreement between the Government of the United States of America and the Government of the People’s Republic of China, January 2020, Article 2.1: ‘Each Party shall ensure fair and equitable market access to persons of the other Party that rely upon intellectual property protection.’ Preamble, TRIPS Agreement.

                                  



106 • a determinant of the legitimacy of traded goods • a tradable service in itself when service delivery entails the licensing of access to the IP • a tradeable good, to the extent that purchase of access to and delivery of the material covered by the IP right is deemed a ‘good’ as such • an avenue for cooperation in research and development, and a bespoke vehicle for transfer of technology107 108 • an asset, the express subject of investment disciplines 109 • a component of and an obstacle to the exercise of human rights, and 110 • as a subsidy.

Physicists cannot directly observe fundamental particles, and can only understand their character by observing their behaviour. In a roughly analogous way, rather than first seeking to define the essence of an IP right in the international trade law and policy space, and then to work out towards its practical manifestations, it may be more illuminating instead to observe how IP rights behave in international commerce, and on that basis accept that – depending on the context and purpose of this observation. Thus IP rights may manifest behaviour as: • • • • • • • 106 107

108

109

110

transacted property traded goods services the subject of market regulation an element of expectations of market access the content of technology transfer or equitable sharing of benefits investment assets

Preamble, and Article 41, TRIPS Agreement. Jayashree Watal and Letitia L. Caminero, ‘Least-Developed Countries, Transfer of Technology and the TRIPS Agreement’, in C. Correa and X. Seuba (eds), Intellectual Property and Development: Understanding the Interfaces, Singapore: Springer, 2019. Lahra Liberti, ‘Intellectual Property Rights in International Investment Agreements: An Overview’, OECD Working Papers on International Investment, 2010/01, OECD Publishing, available at http://dx.doi.org/10.1787/5kmfq1njzl35-en. United Nations, Committee on Economic, Social and Cultural Rights, General Comment No. 17 (2005), The right of everyone to benefit from the protection of the moral and material interests resulting from any scientific, literary or artistic production of which he or she is the author (Article 15, para. 1 (c), of the Covenant), E/C.12/GC/ 1712 January 2006. DS353: United States – Measures Affecting Trade in Large Civil Aircraft – Second Complaint.



 

• subsidies, or • a constraint on trade, technology transfer or fair competition. More succinctly, ‘Old Town Road’, the song discussed in the introduction above, exemplifies how music can manifest the characteristics of distinct categories, depending on the vantage point of the listener, without being solely confined to either category, as the songwriter observed: ‘It’s not one, it’s not the other. It’s both.’111 In other words, rather than insisting on one exclusive categorization, it may be both more practically useful and more empirically sound to observe its behaviour, and to use that understanding as the basis for drawing up new rules or updating old ones. Whatever formal or legal overlay is applied to the new trading arrangements for IP, facilitated by new technologies, it is essential to build an understanding of their character, in part, as trade in IP licences as such, rather than viewing the IP component as an adjunct or ancillary element. The implications for trade policy, and for economic and social development, are profound. In parallel with the TRIPS negotiations, and the contemporaneous invention of the World Wide Web, Paul Romer and other economists were addressing the need for economic growth theory to incorporate knowledge, technological knowhow and human capital as endogenous factors, rather than maintaining it as exogenous to models of growth. The resultant ‘endogenous growth theory’ recalled the importance for human development and economic growth of the institutions and policy settings that promote the development and dissemination of new knowledge. As Romer expressed the analytical task in these terms when accepting the 2018 Nobel Prize for Economics: The human condition emerges from a never-ending contest between the dismal Malthusian economics of objects and the unrealized possibilities of the economics of ideas. For centuries, economists took sides and followed Thomas Malthus. A paper I published finally turned it into a fair fight. Economists no longer have to assume that Malthus wins before exploring the question posed by title I chose for my Nobel lecture: ‘On the possibility of progress’ . . . In my paper112 . . . all I did was make the trivial observation that ideas belong to neither of the standard analytical categories: private goods and public goods . . . I showed that this observation has consequences; big consequences; the biggest possible consequences.

111 112

‘Lil Nas X Talks “Old Town Road” and the Billboard Controversy’, n. 3 supra. Paul Romer, ‘Endogenous Technological Change’ (1990), Journal of Political Economy 98, 5, Part 2: S71–S102.

                                  



The unique characteristics of ideas make material progress possible, but that’s not all. Ideas matter not just for what humans have, but also for how they are. During the Pleistocene, human nature evolved in a Malthusian world of objects. We developed an ugly tendency to split humanity into ‘us’ and ‘them’. A world that also includes ideas justifies a new mindset that treats all humans with the dignity and respect that we offer to ‘us’. It is a world in which we may derive net material benefits from the presence of others.113

Romer has argued that endogenous growth theory enables a shift from the zero-sum mindset that is focused on the inherently rivalrous domain of atoms, and an acceptance of mutual interest in a positive-sum conception of human progress and economic growth – recalling the ‘mutual interest’ that the TRIPS Agreement itself sets as an objective for the IP system in its Article 7. Trade law and trade policy have conventionally viewed the operation of IP rights as, in effect, exogenous. The shifting contours of trade in knowledge – the emergence of widespread trade in IP rights as such, as traded content is detached from chattels, and the integral and essential role of IP licensing in defining and enabling knowledge flows though dispersed global value chains – suggest that trade policy analysis must similarly work to incorporate an understanding of the IP dimension of cross-border commercial exchanges as an integral element of trading relations. This means treating the exchange and licensing of IP rights systematically and effectively as ‘endogenous’ to trade. Such a shift in framing the ‘trade-related aspects’ of IP – turning instead to consider the IP-related aspects of trade – is essential for an accurate empirical picture of trade relations today, especially given the economic significance both of dispersed global value chains and of trade in ‘pure’ IP content as such, particularly in the creative sectors. To work towards such understanding is not an abstract or academic exercise. It is, for instance, essential if developing countries are to benefit from knowledge spillovers from their engagement in global value chains. And it is essential to the creation, evolution and regulations of global digital markets that would enable fair and feasible access to creators across the globe, most critically from developing countries and from least developed countries (LDCs). For instance, the recent ITC report Creative Industries in Rwanda: Digital Paths to Global Markets identifies the 113

Paul Romer, Nobel Lecture: On the Possibility of Progress, 5 February 2019, available at https://paulromer.net/prize/.



 

remarkable opportunities offered to the Rwandan cultural industries that have been opened up by the evolution of digital trading platforms, while identifying ‘common challenges for small businesses exporting creative products and services in developing countries . . . [including] the need for fair revenue sharing with artists; [and] access to, and costs of, operating on global platforms’. For instance, concerning sustainable development of the creative industries: Creative industries involve a complex network of artists, creators, producers, distributors, platforms and intermediaries. Fair sharing of revenue has been a critical issue, as artists and creators are often individuals or small businesses and have less bargaining power against the networks and platforms that often dominate the distribution channels. Digitization is exacerbating this problem, as ‘winner-take-all’ is a common strategy for digital companies.114

This report calls for ‘policymakers and industry players to take action to ensure inclusive and sustainable development of the creative sector’. Romer, in one of his breakthrough papers, underscored the importance of integrating the knowledge component of economic growth into wider policy development in these terms: We will be able to rejoin the ongoing policy debates about . . . the feedback between trade policy and innovation, the scope of protection for intellectual property rights, the links between private firms and universities, the mechanisms for selecting the research areas that receive public support, and the costs and benefits of an explicit government-led technology policy. . . In a developing country like the Philippines, what are the best institutional arrangements for gaining access to the knowledge that already exists in the rest of the world? In a country like the United States, what are the best institutional arrangements for encouraging the production and use of new knowledge?115

Those concluding observations, published the very year the TRIPS Agreement was concluded, seemed to make the assumption – perhaps better empirically grounded at the time – that a stark distinction existed between the interests of a country like the Philippines which is seen in this case entirely as a recipient, as a beneficiary of other people’s technology or knowledge, and a country like the US which is seen as 114

115

International Trade Centre, Creative Industries in Rwanda: Digital Paths to Global Markets, Geneva: ITC, 2019, p. viii, available at www.intracen.org/uploadedFiles/ intracenorg/Content/Publications/Rwanda%20Digital%20exports_final_Low-res.pdf. Paul Romer, ‘The Origins of Endogenous Growth’ (1994), Journal of Economic Perspectives 8, 1: 3–22, p. 22.

                                 



the source of innovation for others. Today’s international knowledge economy is more heterogeneous and less polarised, as many developing countries work their way up the ranks of innovative capacity, strengthen their indigenous innovation capacity and recognize and value traditional knowledge systems. The consequential value is all the more evident, then, of building the IP dimension integrally into trade policy, in a coherent, inclusive and empirically up-to-date manner.

Bibliography AAP StatShot Annual Report for Calendar Year 2018, 21 June 2019. Adjei-Kontoh, Hubert, ‘Lil Nas’ Song was Removed from Billboard for Not Being “Country” Enough. But Who Gets to Decide Categories?’, The Guardian, 2 April 2019. Anell, Lars, ‘Keynote Speech at the TRIPS Symposium, 26 February 2015’, in Jayashree Watal and Antony Taubman, The Making of TRIPS, Geneva: WTO, 2015, pp. 285–291. Aniftos, Rania, ‘Lil Nas X’s “Old Town Road” is the Fastest Song in History to be Certified Diamond by the RIAA’, Billboard, 22 October 2019. Arora, Ashish and Gambardella, Alfonso, ‘Ideas for Rent: An Overview of Markets for Technology’ (June 2010), Industrial and Corporate Change 19, 3: 775–803, available at https://ssrn.com/abstract=1617051 or http://dx.doi.org/dtq022. Bannock, Graham, Davis, E. and Baxter, R. E., The Penguin Dictionary of Economics, London: Penguin, 2004. Chafee, Zechariah, ‘Equitable Servitudes on Chattels’ (June 1928), Harvard Law Review, 41, 8: 945. Chow, Andrew, ‘Lil Nas X Talks “Old Town Road” and the Billboard Controversy’, available at https://time.com/5561466/lil-nas-x-old-town-road-billboard/. Coke, Edward, Institutes of the Laws of England: Containing the Exposition of Many Ancient and Other Statutes, 1628. Department of Foreign Affairs and Trade, Driving Forces on the New Silk Road, Canberra: Department of Foreign Affairs and Trade, 1999. The Economist, ‘The Web’s New Walls’, The Economist, 2 September 2010. Einstein, Albert and Infeld, Leopold, The Evolution of Physics, Cambridge: Cambridge University Press, 1938, available at https://archive.org/details/ evolutionofphysi033254mbp/. Fitzpatrick, David, ‘Negotiating for Hong Kong’, in Jayashree Watal and Antony Taubman, The Making of TRIPS, Geneva: WTO, 2015, pp. 285–291. Frey, Carl Benedikt, Ansar, Atif and Wunsch-Vincent, Sacha, ‘Defining and Measuring the “Market for Brands”: Are Emerging Economies Catching Up?’, Economic Research Working Paper No. 21, 2014.



 

IMF, Balance of Payments and International Investment Position Manual, Washington, DC: International Monetary Fund, 2009, p. 119. International Federation of the Phonographic Industry (IFPI), ‘Global Music Report 2019 // State of the Industry’, available at www.ifpi.org/recordingindustry-in-numbers.php. International Trade Centre, Creative Industries in Rwanda: Digital Paths to Global Markets, Geneva: ITC, 2019. Liberti, Lahra, ‘Intellectual Property Rights in International Investment Agreements: An Overview’, OECD Working Papers on International Investment, January 2010, OECD Publishing, available at http://dx.doi.org/ 10.1787/5kmfq1njzl35-en. Negroponte, Nicholas, ‘Bits and Atoms’ (January 1995), Wired 3, 1. OECD, ‘Intellectual Property Market, in OECD Science, Technology and Industry Outlook’, 2012, OECD Publishing, available at dx.doi.org/10.1787/sti_out look-2012-24-en. Panagariya, A., ‘TRIPS and the WTO: An Uneasy Marriage’, in K. Maskus (ed.), The WTO, Intellectual Property Rights and the Knowledge Economy, Cheltenham: Edward Elgar Publishing, 2004, pp. 42–53. Perzanowski, Aaron and Schultz, Jason, The End of Ownership: Personal Property in the Digital Economy, Cambridge, MA: MIT Press, 2016. Rifkin, Jeremy, The Age of Access, New York: Penguin Putnam, 2000, p. 4. Romer, Paul, ‘Endogenous Technological Change’ (October 1990), Journal of Political Economy 98, 5, Part 2: S71–S102. Romer, Paul, ‘The Origins of Endogenous Growth’ (1994), Journal of Economic Perspectives 8, 1: 3–22. Romer, Paul, Nobel Lecture: On the Possibility of Progress, 5 February 2018, available at https://paulromer.net/prize/. Smith, Adam, An Inquiry into the Nature and Causes of the Wealth of Nations, Henry Frowde (ed.), Oxford: Oxford University Press, 1909 (1776). Taubman, Antony, ‘Nobility of Interpretation: Equity, Retrospectivity, and Collectivity in Implementing New Norms for Performers’ Rights’ (2005), Journal of Intellectual Property Law 12, 351, available at https:// digitalcommons.law.uga.edu/jipl/vol12/iss2/2. Taubman, Antony, ‘TRIPS Encounters the Internet: An Analogue Treaty in a Digital Age, or the First Trade 2.0 Agreement?’, in Mira Burri and Thomas Cottier (eds), Trade Governance in the Digital Age, New York: Cambridge University Press, 2015, p. 314. United Nations, Committee on Economic, Social and Cultural Rights, General Comment No. 17 (2005), ‘The right of everyone to benefit from the protection of the moral and material interests resulting from any scientific, literary or artistic production of which he or she is the author’ (Article 15, para. 1 (c), of the Covenant), E/C.12/GC/1712 January 2006.

                                 



United Nations, International Merchandise Trade Statistics: Concepts and Definitions 2010 (IMTS 2010), New York: United Nations, 2010, p. 15. United Nations, ‘International Merchandise Trade Statistics: Concepts and Definitions’ 2010, ST/ESA/STAT/SER.M/52/Rev.3 Statistical Papers Series M No. 52, 12. US Copyright Office, ‘A Report of the Register of Copzrights Pursuant § 104 of the Digital Millennium Copyright Act, 2001’, available at www.copyright.gov/ reports/studies/dmca/sec-104-report-vol-1.pdf. Van Welsum, D., Overmeer, W. and van Ark, B., ‘Unlocking the ICT Growth Potential in Europe: Enabling People and Businesses’, The Conference Board for the European Commission, Brussels: European Commission, 2013. Watal, Jayashree and Caminero, Letiti, ‘Transfer of Technology and the TRIPS Agreement’, in C. Correa and X. Seuba (eds), Intellectual Property and Development: Understanding the Interfaces, Singapore: Springer, 2019. WIPO, ‘The Global Publishing Industry in 2016: A Pilot Survey’, 2016, available at www.wipo_ipa_pilotsurvey. WIPO, The Global Publishing Industry in 2018, Geneva: World Intellectual Property Organization, 2020. WTO, Trade Policy, The WTO and the Digital Economy, Communication from Canada, Chile, Colombia, Côte d’Ivoire, the European Union, the Republic of Korea, Mexico, Montenegro, Paraguay, Singapore and Turkey JOB/GC/ 116, JOB/CTG/4, JOB/SERV/248, JOB/IP/21, JOB/DEV/42 (13 January 2017). WTO, World Trade Report 2018: The Future of World Trade: How Digital Technologies are Transforming Global Commerce, Geneva: WTO, 2018. Wunsch-Vincent, Sasha, ‘The Internet, Cross-Border Trade in Services, and the Gats – Lessons from US-Gambling’ (2006), World Trade Review 5, 3: 319–356, p. 323. Zittrain, Jonathan, ‘The Generative Internet’ (1974), Harvard Law Review 119, 7.

3 How Digitization is Transforming Trade           ,                         *

Abstract This chapter looks at the ways new digital trends and technologies have stimulated the information component of services trade and consequently enhanced trade in goods and services that embody knowledge. It illustrates the role that ICTs have taken on as conduits for digital and digitallyenabled trade and describes recent digital developments such as cloud services, large scale data analytics, Internet of things, artificial intelligence, robotics, and three-dimensional printing. It presents the relevant trade data and then covers, in broad strokes, the landscape of policy challenges that governments confront as they seek to adapt, and examples of ways trade negotiators have begun to shape new legal frameworks. The rise of such technologies and corresponding growth in services that trade internationally by means of the global movements of information has led policymakers to believe a new or enhanced legal regime could be needed. Not always well understood, digital trends can lead to exaggerated fears and cataclysmic predictions. Over-reaction to these transformations can hold the risk of disproportionate policy responses, which may harm not only trade, but also the sharing of knowledge across borders. One significant challenge, the article nevertheless raises, is that international trade rules are founded on the premise that national governments can implement them at or within their borders. Yet, many of the policy and legal responses that arise from an unanticipated shift in services trade from commercial presence to crossborder supply, have a variety of interjurisdictional consequences. Therefore, whatever trade rules may be employed, existing or new, the article concludes that enhanced efforts at collaboration among governments will be needed to complement and coordinate national initiatives.

* The views expressed are the authors’ own and should not be attributed to any WTO member or the WTO Secretariat. The authors would like to extend thanks to Joscelyn Magdeleine, Justine Lan and Kian Cassehgari Posada for valuable input and assistance.



      



Introduction – Information in Trade and Trade in Information The expression ‘knowledge is power’, attributed to Sir Francis Bacon, is a point worth applying to trade. Knowledge embedded in goods and services arises from innovation, expertise and creativity. It creates value and motivates society to trade. In short, knowledge is a power that drives markets. Digitization is transforming trade. At the core of such a transformation are the various information and communication technology (ICT) products that provide the basic infrastructure for digital supply. Technological knowledge is intrinsic to those ICT goods and services. Digitization has, in particular, vastly increased opportunities for trade in information or knowledge-intensive services. Services trade, for its part, has long been characterized as knowledge and information intensive, and, as a result, has proven easily amenable to trade by electronic means. So-called knowledge services range from the professions and research and development to education and consultancy. Once information at the core of such services is digitized, it can be supplied via digital media in both domestic and international settings. Information services such as databases, news agency services, or media platforms can now offer the desired content online directly to consumers, wherever they may be located. In addition, ICT services arising from technological knowhow, such as telecommunication and computer services are not only key to facilitating further trade in knowledge and information-intensive services, but are also traded themselves, not only across borders but also by means of establishment by foreign suppliers. Moreover, goods embody knowledge services in the form of, for example, design services or software programming that form an important part of the value chain of their production. More recently, as the services component of goods production has grown, the incidence of knowledge services associated with merchandise goods has also increased. Manufacturing is increasingly reliant on service inputs and progressively supplying services alongside goods, in a process referred to as ‘servicification’. Together with the deployment of ICT tools, this allows producers to become better informed of and more responsive to their customers’ needs, which in turn facilitates innovation and sustainability, and to maintain mutually beneficial long-term relations. Consequently, greater digitization and the further evolution of ICTs have increased the contribution of intellectual property and other forms of knowledge to trade. WIPO estimates that, on average, 30.4 per cent of the production value of global value chains is generated by intangible



  ,      

capital in the form of technology, design and brand value, as well as workers’ skills and managerial knowhow.1 This chapter looks at the ways new digital trends and technologies have stimulated the information component of services trade and consequently enhanced trade in goods and services that embody knowledge. It first illustrates the role that ICTs have taken on as conduits for digital and digitally-enabled trade.2 It briefly describes recent digital developments, namely the cloud, data analytics, (so-called ‘big data’), Internet of Things (IoT), aka machine-to-machine (M2M) communications, artificial intelligence, robotics, and three-dimensional (3D) printing (aka additive manufacturing). It covers, in broad strokes, the landscape of policy challenges that governments confront as they seek to adapt. Finally, the chapter provides illustrations of ways negotiators of trade rules have begun to shape new legal frameworks via regional trade agreements (RTAs) that elaborate upon, and in some respects extend beyond, existing multilateral trade rules. While the focus of this chapter is primarily on trade in services, some of the implications for trade in merchandise are also noted.

ICT – the Conduits for Digital Trade From Physical Proximity to Phone and Fax Early research on trade in services, including studies by trade officials laying the groundwork for services trade negotiations in the late 1980s, revealed that services were commonly traded via physical proximity of the supplier to the consumer. Among the reasons for this feature was frequent customization, or personalization, for example, of legal advice or accountancy services to clients. This meant that location of a firm or of an individual service provider in the country where the services were to be offered was necessary to facilitate direct interaction and negotiation with clients on the desired features. Another reason for physical proximity was the degree to which a number of services were not considered storable, given their generally intangible nature, and, as such, not easily 1

2

WIPO, World Intellectual Property Report 2017: Intangible Capital in Global Value Chains, Geneva: WIPO, 2017, p. 11. This chapter uses interchangeably terms such as ‘digitally-enabled trade’, ‘digitallyimpacted trade’, ‘online trade’ and similar. As the equivalent term in the WTO context is ‘e-commerce’, as defined by the Work Programme on Electronic Commerce, the latter is also used, in particular with reference to discussions in the WTO.

      



transmittable across borders. Services, such as a consultant’s advice, if not provided in person, could be conveyed by a phone call or by mail. In essence, services trade traditionally relied on personal interaction or was embedded in physical carrier mediums. Both of these options incurred costs in terms of setting up a commercial establishment, travelling to the consumers’ market or transporting items by physical means. Indeed, the customization and non-storability of services led to the creation of a new paradigm for trade. The General Agreement on Trade in Services (GATS) re-engineered trade disciplines to cover services traded not only across borders but rather by four different modes of supply. The first concerned trade across borders (also known as crossborder supply, or mode 1, in GATS parlance), but the other three modes of supply extended the definition of trade to the consumer moving into physical proximity with the supplier (consumption abroad, mode 2), the establishment of a business abroad (commercial presence, mode 3) and an individual supplier or the foreign employee of an enterprise engaged in services physically visiting the client’s country (movement of natural persons, mode 4). Services supplied by phone and fax were acknowledged as means to cross-border trade.3 However, in a gesture towards the evolving digital technologies, the GATS Annex on Telecommunications recognized the sector’s ‘dual role as a distinct sector of economic activity and as the underlying transport means for other economic activities’, and subjected telecommunications to specific trade disciplines. Such disciplines referred to trade that was facilitated, even then, by communications networks used by companies to supply services. Market reforms inspired further innovation in telecommunications, such as a surge in mobile services and broadband Internet technology.

From Telephony to Data Networks and Internet Early networks for corporate communications managed corporate affairs across global subsidiaries and affiliates as well as with global customers and suppliers. These drove the emerging trade in so-called ‘value-added 3

As cited in the first and subsequently revised guidelines for scheduling of market access commitments in services (documents MTN.GNS/W/164, Scheduling of Initial Commitments in Trade in Services: Explanatory Note, dated 3 September 1993, and MTN.GNS/W/164/Add.1, dated 30 November 1993 and revised as S/L/92, adopted by the Council for Trade in Services on 23 March 2001).



  ,      

networks services’, allowing firms to outsource, instead of self-service, their communication needs. They offered their clients services such as email, database and data processing services. Commercialization of the Internet began at the turn of the 1990s, primarily as vehicle for email and some relatively limited information services. It was accessed on lines dialled-in over the public telephony networks. Despite slow speeds and the relatively high cost to consumers, Internet uptake nevertheless grew at rates faster than any previously encountered technological innovations. Growth was so phenomenal that the expression was coined that an Internet year was equivalent to three months.4 Even then, Internet growth was already accompanied by concerns about policy and legal frameworks keeping pace with its development and the transformation in human activity. Internet access as we know it today, however, was not possible until the introduction of the World Wide Web in 1991 and easily accessible web browsers (Mosaic in 1992 and Netscape in 1994) permitting more accessible use and more sophisticated websites with greater volumes of information. Speeds and capacity (bandwidth) to more easily permit business applications, such as online ordering, and encryption to permit customer accounts, purchase transactions and payment mechanisms, first arose in the early 2000s. Broadband Internet, both fixed and mobile, is growing in the developed world, and is still emerging in the developing world where penetration rates remain low. With the advent of newer technologies, not only the Internet, but, subsequently, platforms and other services overlaid on a better and faster Internet had a multiplier effect on economic activity, leading to what some observers see as digital disruption and to a redefinition of commercial models and trade patterns. The feasibility of offering some of the most popular services over the Internet today, e.g. music and film distribution, had to await not only higher capacity bandwidth, but also the emergence of new business models in the entertainment industry. These industries had long considered the Internet mainly to be a vehicle for piracy or revenue loss. The rise of mobile services as the telephony of popular choice, and now of mobile Internet, was one of the single most unanticipated phenomena in communications – not only because the reduced cost and mobile Internet technology was not foreseen by many observers, but also because it contributed to 4

Attributed to Tim Berners-Lee, chief inventor of the Web technologies, in a 1996 interview in the WWW Journal.

      



vastly expanding the penetration of communications technologies in developing countries.

Evolving High-Capacity Internet The advance that truly makes digital business and trade a reality is higher speed, higher-capacity Internet, characterized by more powerful bandwidth. This was accomplished by, for example, fibreoptic cable rollout by telephone operators, as well as the introduction of Internet by way of the upgrade of networks previously dedicated to cable television. In mobile telephony, the highest bandwidth offerings currently possible, known as fifth generation, or 5G, are just beginning to be introduced in most countries. Other advances discussed below and heralded today as leading to a transformation and redefinition of trade ultimately rely on highcapacity Internet for their existence and commercial viability. The expansion of knowledge in the form of technological knowhow has led ICT infrastructure services to become vehicles of ICT-enabled trade for other knowledge and information-intensive services. These trends have impacted trade more generally, in a number of ways: • First, they have enhanced the tradability of services, thereby expanding export opportunities for both developed and developing countries. It is increasingly easier and less costly to export services across borders, either as final products or as intermediate products for the production of another service or a good. For suppliers across a range of services, cross-border supply becomes either a new – or more interesting – alternative, or a complement to the supply of services through a commercial presence abroad (mode 3) or the temporary movement of persons (mode 4). • Second, the increased performance of ICT network services, in combination with innovations in logistics, have enabled an enormous expansion of global value chains. Information-intensive services, such as computer, research and development, advertising, telecommunications, and financial and professional services serve as enablers of global value chains. • Third, ICT services underpin flows of communications in the form of data flows across borders, which have skyrocketed in recent years. Cross-border data flows, boosted by basic and value-added telecom services, such as data processing and storage via large capacity cloud technology, allow companies not only to sell their goods and services,



  ,      

but also to coordinate their logistics and the activities of their subsidiaries and partner offices across the globe.5 • Fourth, ICT services, and more specifically the Internet, constitute the backbone for key pillars of e-commerce such as online retail and wholesale trade, whether cross-border or domestic. Indeed, without increased capacity and speed, and the lower communication costs brought about by improvements in telecom and computer services, the sale of goods online as it stands today, including inventory management, would not be possible.6 The future holds much more potential than we can imagine today. This is evident in the still low uptake of broadband Internet, particularly in the developing world, and in the even lower uptake of mobile Internet. Although twice as many mobile-broadband subscriptions exist per 100 inhabitants in developed countries than in developing countries, and four times as many in developed countries than in least-developed countries (LDCs), the growth rate of mobile-broadband subscriptions was the highest in LDCs and developing countries, at over 50 per cent and 30 per cent, respectively, from 2016 to 2017. Over the same period, the growth of fixed-broadband subscriptions climbed by over 40 per cent and 10 per cent in LDCs and developing countries, respectively, as compared to less than 5 per cent in developed countries. Mobile broadband is also increasingly more affordable worldwide. Mobile-broadband prices fell from 32.4 per cent to 14.1 per cent of gross national income (GNI) per capita in LDCs, from 11.6 per cent to 6 per cent in developing countries, and from 1 per cent to 0.7 per cent in developed countries, from 2013 to 2016.7

Digital Trends to Watch The six digital trends taking shape today that appear as most transformative are cloud technologies, the Internet of Things (IoT), large-scale data analytics (also known as big data), artificial intelligence, robotics and three-dimensional (3D) printing. 5

6

7

L. Tuthill, ‘Cross-border Data Flows: What Role for Trade Rules?’, in P. Sauvé and M. Roy (eds), Research Handbook on Trade in Services, Cheltenham: Edward Elgar Publishing, 2016. M. Roy, ‘The Contribution of Services Trade Policies to Connectivity in The Context of Aid for Trade’, WTO Staff Working Paper, ERSD-2017-12, 2017. International Telecommunications Union (ITU), 2017 ‘ICT Facts and Figures 2017’, available at www.itu.int/en/ITU-D/Statistics/Pages/facts/default.aspx.

      



These developments, although diverse, are characterized, first and foremost, by significant overlap and synergies. Second, some represent new commercial models for super-charged versions of relatively longexistent types of computer services. And last, but not least, the terms used for these developments do not represent services, in themselves, but rather are generic expressions that encompass families of evolving technologies that can be deployed to supply a wide array of services, most with which we have long been familiar. To illustrate these points, ‘cloud services’, for example, are essentially a commercially generated term that denotes configurations of networks and equipment by which service suppliers can offer high-capacity storage, data processing and delivery options to their customers. Large-scale data analytics was born out of the coming of age of the cloud’s massive storage and computing capacity. Robotics and IoT can only be realized in conjunction with data network and transmission services and are subsequently used to supply a vast array of services offerings. Artificial intelligence and robotics result in large part from sophisticated algorithms and software. IoT, for its part, creates data that becomes fodder for more powerful data processing capacities. As it becomes increasingly commercialized, they will be deployed via chips or intelligent, software-embedded devices performing many different functions and services. Most often, these technologies will rely on wire-based technologies, such as satellite or mobile links, connected to data networks, typically the Internet, permitting information to be transmitted. However, the data only become commercially viable once interpreted by means of data processing methodologies. Finally, 3D printing results from an extension of advanced software and design services into the realm of equipment capable of producing replicas of physical objects. While such software may be inserted into printing equipment via a physical cartridge, again it can also be linked to ICTs when it is transmitted virtually to network-enabled printers.

Impacts – Redefining Trade? The most direct impact of this family of technological developments is to expand the universe, and, more specifically, the amplitude of services that can be traded. At the same time, they are also not without implications for goods production and trade. The ease by which these digital trends can augment global services trade is directly attributable to their being, by their



  ,      

very nature, networked technologies. They are conveyed by networks, like the Internet, which, in turn, are global in nature. The economic activities arising from these developments will be made available to other businesses by ‘next-generation’ digital service suppliers. By extension, given the information intensity of most services, a great many customers of these digital services will, themselves, be suppliers of services. Moreover, the variety of services that can ultimately be supplied in this chain is boundless. Another transformative implication of the networked nature of these trends is that the potential for trade via remote supply will be even further enhanced, and at a reduced cost, permitting even greater segmentation of value chains across borders than is the case today. Such segmentation of production, for example, in the research and development sector, IT services and business support services has already shown signs of increasing the participation of developing countries in global trade. Yet, many of the digital trends and associated challenges can lend themselves to exaggeration, outsized fear for the future and cataclysmic predictions. A case in point is the early notion that e-commerce would lead to the death of intermediaries, with consumers ordering directly from factories. Instead, a new breed of online intermediaries evolved.

Economic Significance of Digitally-Enabled Trade Among the most significant impacts of digital technologies is the reduction of trade costs and greater ease of market entry. The trade data appear to support these implications. Advances in information and communications technology are transforming the tradability of services, increasing the growth of services supplied electronically cross-border, across a wide range of sectors, from medical, to educational, financial, audiovisual or professional services. This is reflected, in the first instance, in the growing relative importance of services in total trade. Measured on a balance-of-payments (BoP) basis, trade in services now accounts for 23 per cent of total trade in goods and services, compared to 19 per cent in 1995. The strong growth of trade in services is largely a result of the Internet revolution.8 8

BoP statistics do not cover all four modes of services supply defined by the GATS, as it excludes mode 3. Among other modes of supply, cross-border trade (mode 1) is estimated to account for 27 per cent of world trade in services, compared with 15 per cent for mode 2, 5 per cent for mode 4, and 53 per cent for mode 3 (WTO World Statistical Review 2018, WTO, Geneva, available at https://doi.org/10.30875/0ab3aa40-en).

      



Moreover, technological advances and increased tradability have led to significant change in the composition of trade in services. ‘Other commercial services’, which lend themselves more easily to digitisation, have increased significantly and now account for 54 per cent (in 2017) of global trade in commercial services and include some of the most dynamic components of world trade today, compared with less than 40 per cent in 1995, and 47 per cent in 2005. By contrast, the relative share of travel and transport in BoP statistics has diminished considerably. Not surprisingly, services that have experienced strong growth consist of services that can be supplied electronically and that have benefitted significantly from the increased efficiency of digital networks. These are also services intensive in digitized information and technological knowledge. Between 2005 and 2017, exports of the fastest-growing subcategory, ‘telecommunications, computer, and information services’ increased at an annual average rate of 13.3 per cent; exports of these services now account for 10 per cent of world exports, compared with 7.8 per cent in 2005. Most of the trade in this category pertained to computer services, which had an annual growth rate of 17.6 per cent over the same period. Exports of what may be referred to as ‘IP services’ (i.e. charges for the use of intellectual property not included elsewhere) exhibited an average annual growth rate of 11 per cent over the same period. With 10 per cent growth in 2017, this was the fastest-growing service category. This knowledge-intensive category includes charges for the use of proprietary rights (e.g. patents, trademarks, copyrights, franchises, and rights arising from research and development) as well as charges for the authorized reproduction and/or distribution (through licensing agreements) of produced software originals, as well as fees and charges for the authorized reproduction/distribution, through licensing agreements, of produced audiovisual originals or prototypes (for example, cinematographic works and sound recordings) and related rights pertaining to recordings of live performances and radio, television, cable and satellite broadcasts.9 A vast array of information-intensive services is captured in the category of ‘other business services’, which accounted for 23 per cent of total services exports in 2017. These services, which grew at an average 9

United Nations Statistical Commission (2010), Manual on Statistics of International Trade in Services (MSITS 2010) 2010, United Nations publication, 2010, available at https://unstats.un.org/unsd/publication/Seriesm/seriesM_86Rev1e.pdf.



  ,      

annual pace of 11 per cent between 2005 and 2017, include, for example, research and development services as well as professional, advertising and management consulting services. Many developing countries are taking greater advantage of direct export opportunities offered by new technologies enabling the digital supply of services. Their share of global trade in services has risen from 28 per cent in 2005 to 34 per cent in 2017. Exports of developing countries are concentrated in ‘other commercial services’ (44 per cent in 2017), in contrast to transport (20 per cent), travel (33 per cent), and goods-related services (3 per cent). With 13.7 per cent of world exports, India ranked as a top global exporter of computer services in 2017, second only to the European Union.

Potential for Digital Disruption? While, on the one hand, changing trade patterns are seen as a boost to developing economies, this transition may, on the other hand, be perceived as disruptive of the status quo in industrial economies. Advanced economies could, in the not too distant past, count on having a comparative advantage in technology, information and skills needed for a knowledge economy. Certainly, economic disruption can occur through the demise of certain economic actors, but also through the creation of new ones. However, it is possible that the shift of capital, skills and technologies required for growth and comparative advantage in trade could well be smoother for less advanced countries without legacy technologies and industries, permitting them to transition directly to new economic models. Many developing countries, have, for example, moved directly into mobile communications, without having built fixed telecommunication networks, and they have secured mobile banking and payments methods for citizens that never had a bank account. Meanwhile, such services have been much slower to take hold in developed countries where legacy telecommunications and banking services were widely available. The implications of technological development for manufacturing are most apparent in robotics and 3D printing, which lead to concomitant implications for goods trade. While robotics can make possible the remote provision of services, such as surgical operations, its most transformative applications have, to date, taken hold on factory floors in the manufacturing industry. By lowering the costs of production, competitive advantage can be enhanced, yet the transition from labour-intensive to

      



capital-intensive manufacturing can be disruptive, particularly in the job market and in developing countries. For 3D printing, compared with robotics, commercial applications are still in their infancy. Nevertheless, the potential for 3D applications to transform manufacturing industries is vast, if not sometimes overstated, at least in the near-term outlook.

Twenty-First-Century Trade Rules? Challenges The Internet blurs geographic boundaries, raising issues about the proper application of private international law, including jurisdiction, applicable law, and the recognition and enforcement of foreign judicial rulings, that some domestic and international trade rules may not be well equipped to address. Even international trade rules, for example, are founded on the premise that national governments will implement them at their borders or within their own territories. Some of the policy and legal questions that arise also stem from an unanticipated shift in services trade away from commercial presence to cross-border supply. It is, moreover, incumbent on trade negotiators to discuss not only aspects of digital technologies that directly impact trade, but also to address issues less directly related to trade that touch on other public policy concerns. For example, a rise in cross-border data flows is an intrinsic and essential result of trade conducted via new technologies. Yet, controversies underpin policy concerns linked to a great many digital trade flows and transactions, most notably those with a bearing on so-called non-trade matters, such as privacy, cybersecurity and protection against fraud, hacking and other cybercrimes. The rise of cloud computing, IoT, data analytics and other services that internationally transfer and store data has led some policymakers to believe a new or enhanced legal regime could be needed to handle data moved across borders. With more companies building or outsourcing data centres around the world, questions may arise about the relevance of server location in determining jurisdiction in certain matters, such as judicial proceedings, or even WTO proceedings. The jurisdictional issues can best be explained by an ongoing case involving Microsoft. It came about after a 2013 drug trafficking investigation during which the FBI issued a Stored Communications Act (SCA) warrant for emails that belonged to a US citizen that Microsoft had stored on a server in Ireland. Microsoft argued that it could not turn over the data stored



  ,      

outside the US and that the SCA did not cover such cases. The core of the issue is whether a government can compel a service provider to hand over data, even if this is stored in another country. Competition issues may also arise due to network effects. It is easiest to illustrate this effect by looking at social media, where the more users sign up on, for instance, Facebook, the more attractive it is for others to sign up, and the more attractive it is for advertisers to position ads on that platform. In effect, the value of a product or a service increases as more users subscribe to it.

The Existing Framework – WTO Rules The General Agreement on Trade in Services The GATS makes no distinction regarding different technological means through which a service may be supplied. The supply of services through electronic means is covered by the Agreement in the same way as all other means of supply. This has been confirmed by WTO jurisprudence. For example, in US – Gambling,10 the Panel concluded that: mode 1 includes all means of delivery. We are of the view that when a Member inscribes the word ‘None’ in the market access column of its schedule for mode 1, it commits itself not to maintain measures which prohibit the use of one, several or all means of delivery under mode 1 in a committed sector or sub-sector. This is especially so in sectors and subsectors where cross-border supply is affected essentially if not exclusively through the Internet.11

In that context, online trade falls within the scope of GATS, and therefore trade restrictions, as well as domestic regulations affecting electronic supply of services, are subject to GATS obligations, disciplines and commitments.

10

11

Appellate Body Report, United States – Measures Affecting the Cross-Border Supply of Gambling and Betting Services, WT/DS285/AB/R, adopted 20 April 2005, DSR 2005:XII, p. 5663 (and Corr.1, DSR 2006:XII, p. 5475); Panel Report, United States – Measures Affecting the Cross-Border Supply of Gambling and Betting Services, WT/DS285/R, adopted 20 April 2005, as modified by Appellate Body Report WT/DS285/AB/R, DSR 2005:XII, p. 5797. Panel Report, United States – Measures Affecting the Cross-Border Supply of Gambling and Betting Services, at para. 6.287. A ‘none’ inscription in a schedule signifies that no restrictions are applicable.

      



As most GATS disciplines apply only to committed services, the most advantageous conditions for the digital supply of information-intensive services are achieved when relevant commitments exist and when those are as open as possible. The fact that most commitments date from negotiations concluded over 20 years ago represents the single most important gap in the WTO coverage of electronic commerce12 in services. Updating these commitments would be possible, should WTO members decide to do so, given the considerable levels of liberalization in place that has not been bound in GATS schedules. However, in order to most effectively undertake commitments that reflect the current realities of global online trade, some services classification issues may also need to be addressed. But many online services available today are not explicitly mentioned by their common or commercial names, in the 1991 classification list currently used for GATS commitments. Examples include not only various ‘cloud’ services, search engines, web hosting or mobile applications, but also many information services such as social networks and blogs or media services such as magazines, news or other virtual information platforms. Are some of these entirely new services, or do they relate to existing services classification if interpreted in generic terms? Are some of these hybrids combining various services sub-categories together in a manner that can ultimately be linked to an existing classification? To some extent, WTO dispute settlement panels have already faced such issues. In China – Publications and Audiovisual Products,13 for example, the Panel and the Appellate Body noted that the service entry of ‘Sound Recording Distribution Services’, in the context of China’s GATS Schedule, did not expressly exclude any particular form of distribution, i.e. it did not indicate that the commitment was limited to the physical distribution of sound recordings embedded in physical media, and thus concluded that the entry also covered electronic

12

13

As will be discussed below, e-commerce is broadly defined in the WTO, so as to include all forms of digital trade online as well as digital inputs to goods or services that are ultimately delivered physically. Appellate Body report, China – Measures Affecting Trading Rights and Distribution Services for Certain Publications and Audiovisual Entertainment Products, WT/DS363/ AB/R, adopted 19 January 2010, DSR 2010:I, p. 3; Panel Report, China – Measures Affecting Trading Rights and Distribution Services for Certain Publications and Audiovisual Entertainment Products, WT/DS363/R and Corr.1, adopted 19 January 2010, as modified by Appellate Body Report WT/DS363/AB/R, DSR 2010:II, p. 261.



  ,      

distribution.14,15 Also, in the case on China – Electronic Payment Services,16 the Panel dealt with an integrated set of services that together made up an electronic payments system. It ultimately found that the legal requirements China maintained on electronic payment networks for payment card transactions and the suppliers of those services (modes 1 and 3) were inconsistent with its commitments under the GATS. Electronic networks have also made possible digitally-enabled trade that cuts across a great many sectors of activity. Not all of these sectors would be considered ‘knowledge services’, but a great many are. Some of these are already subject to GATS commitments by WTO members. For example, as Annex 1 indicates, the main subsectors of the computer services sector, which entails very high IP content, have been committed in 80 to 89 schedules. Between 72 per cent and 79 per cent of these commitments are unrestricted (i.e. without limitations) for crossborder supply. In terms of enabling services, subsectors of telecommunication services, also a quintessentially electronic service, have also attracted a relatively high number of GATS commitments, though those tend to be subject, in comparison with computer services, not only to more limitations in mode 1, but also mode 3. For example, 65 per cent of commitments on data transmission are subject to limitations or are ‘unbound’ (i.e. uncommitted) with respect to mode 1, and this figure reaches 78 per cent for mode 3. A total of 103 WTO members have committed in whole or in part, in their schedule of specific commitments, to the telecom Reference Paper’s regulatory principles, which are aimed at fostering competition, transparency and predictability in telecom markets and which have led to the extensive innovation in digital infrastructure and services witnessed

14

15

16

Panel report, China – Measures Affecting Trading Rights and Distribution Services for Certain Publications and Audiovisual Entertainment Products, at para. 7.1265, as upheld by the Appellate Body report, China – Measures Affecting Trading Rights and Distribution Services for Certain Publications and Audiovisual Entertainment Products, at paras 375–413. The Panel also noted that ‘digital or electronic distribution’ of sound recordings includes distribution both through the Internet and by other electronic means. Panel report, China – Measures Affecting Trading Rights and Distribution Services for Certain Publications and Audiovisual Entertainment Products, at paras 7.1148–7.1153. Panel Report, China – Certain Measures Affecting Electronic Payment Services, WT/ DS413/R and Add.1, adopted 31 August 2012, DSR 2012:X, p. 5305.

      



today.17 Moreover, the access and use disciplines of the Annex on Telecommunications form an integral part of the GATS. The applicability of the Annex and Reference Paper were clarified in the findings of the GATS case on Mexico – Telecoms.18

The GATT and Associated Agreements Regarding trade in goods, even if not specifically mentioned in any provision, goods sold or marketed by electronic means but delivered physically across borders are without question subject to the existing WTO rules. These rules apply also to knowledge-embodying digital products that are internationally traded by means of a physical support, such as data or software recorded in a DVD. The GATT and other goods texts, such as the agreements on technical barriers to trade, trade-related investment measures, rules of origin, customs valuation, all apply to goods that are physically delivered even if traded electronically. Like the GATS, these agreements are drafted in a technology neutral manner. The more recently concluded Trade Facilitation Agreement (TFA) provides for electronic ways of expediting and facilitating government procedures by, inter alia, encouraging members to accept paper or electronic copies of supporting documents required for import, export or transit formalities. The Information Technology Agreement (ITA) also has important and positive implications for electronic trade. By eliminating tariffs on IT products that are important components of the digital infrastructure, it makes them more affordable and readily available. Possible Multilateral Rules through the Prism of Regional Trade Agreements? Conduits for Digital Trade – Telecommunications Rule-Making Compared to the GATS, in the RTAs, WTO members have expanded their commitments for the telecom sector, as well as extended the regulatory disciplines beyond those in the Reference Paper and 17

18

The Reference Paper on Basic Telecommunications (WTO Job document 2104, dated 24 April 1996). Panel report, Mexico – Measures Affecting Telecommunications Services, WT/DS204/R, adopted 1 June 2004, DSR 2004: IV, p. 1537.

   ,      

Telecom Annex.19 The majority of RTAs cover more services and provide a greater degree of market openness and foreign ownership than that provided for in GATS commitments. In some cases, WTO members with no GATS commitments in basic telecommunication services made RTA commitments on the sector, with no limitations. The most important reason for this is that governments have continued sector reforms and were able to commit to these in RTAs at a time when WTO negotiations are blocked. Out of the 307 notified and non-notified RTAs (as of March 2018), 83 contain a stand-alone chapter on telecommunications with telecoms regulatory principles. RTAs between developing and developed countries make up 51 per cent of RTAs with a stand-alone chapter on telecoms, while RTAs between developing countries as well as RTAs between developed countries represent 44 per cent and 5 per cent respectively. The four participants most frequently involved in RTAs with a telecom component are Chile, the United States, Singapore and the European Union. A number of RTAs broaden the scope of the Reference Paper obligations beyond basic telecommunications, to cover all telecom services, thereby promoting fair and transparent competition for all forms of telecommunications. The provisions also add clarity to the scope of application of the WTO regulatory disciplines, by covering explicitly many services and situations not spelled out in the GATS disciplines, and by extending benefits to value-added telecom services that do not fall within the scope of the GATS Reference Paper. Another RTA expansion on GATS provisions includes requirements for major suppliers to provide leased circuit services at ‘capacity-based, cost-oriented prices’. Leased circuits are used by many digital service providers to create virtual networks to deliver their services. In the GATS Telecom Annex, the standard is vaguer, requiring ‘reasonable and non-discriminatory’ terms and conditions. Further, some RTAs contain new provisions dealing with regulatory issues not addressed – at least, not explicitly – in GATS disciplines. In addition, considering that broadband technology is increasingly both mobile and global, a number of RTAs contain provisions encouraging cooperation to foster fair and transparent pricing with respect to international mobile roaming rates and technological alternatives to roaming services. Such provisions were intended to enable 19

This subsection draws in part on material prepared on ‘telecommunications regulatory issues’ for the World Trade Report 2018, Geneva: WTO, 2018, pp. 182–184.

      



end-users, including business users, to use their mobile handset or other device for voice, data or messaging services while outside their territory at reasonable cost.

Experiments in E-Commerce Rule-Making E-commerce provisions have become an increasingly common feature of RTAs over the past 10 years. Out of 275 RTAs notified to the WTO (as of May 2017), 75 contain e-commerce chapters or provisions. Although this is a relatively small share, accounting for just over 20 per cent of all notified RTAs, the number is growing. Membership of these RTAs is increasingly more diverse, more recently involving agreements exclusively between developing countries. The e-commerce chapters in RTAs normally deal with similar themes. Some address issues, such as cooperation, transparency and nondiscrimination, which already exist in these agreements, but whose scope is extended to explicitly cover e‑commerce. Some RTAs include provisions addressing topics more specific to e-commerce, such as information flows, electronic authentication, paperless trade, and consumer and data protection. In terms of information flows, the ability to transfer data across borders is essential not only to electronic commerce, but also to trade in knowledge that is digitized. The GATS Annex on Telecommunications provides that ‘each Member shall ensure that service suppliers of any other Member may use public telecommunications transport networks and services for the movement of information within and across borders’. While this provision is relevant, it ensures suppliers’ rights to access the underlying telecommunications used as a means to transfer information, but only applies to the extent that the suppliers involved are service suppliers taking advantage of GATS commitments. Of the RTAs with an e-commerce chapter, only a third deal with cross-border information flows. Generally, RTAs contain soft provisions on cooperation or best endeavour provisions to avoid unnecessary barriers to electronic information flows. Very few RTAs, to date, have binding commitments. The CPTPP and the revised Singapore– Australia Agreement, for instance, provide that the parties shall allow the cross-border transfer of information by electronic means, including personal information, when this activity is for the conduct of the business of a covered person. However, they also allow the parties to adopt measures that are inconsistent with this obligation to achieve a legitimate

   ,      

public policy objective, provided that the measures are not applied in an arbitrary or unjustifiable manner and do not impose restrictions greater than what is required to achieve the objective concerned. Requirements to locate data centres, servers or other computing facilities within one’s territory are a form of data flow restriction that may, at most, make data transfer impossible, or, at the least, require duplicate facilities to be established that might not otherwise be needed from a commercial standpoint. Such requirements, depending on how they are designed and implemented, might, under some circumstances, already be disciplined by specific commitments under the GATS. Nonetheless, explicit disciplines on localization are found only in very recently negotiated RTAs, such as the CPTPP, the Japan–Mongolia Agreement and the third revision of the Singapore–Australia Agreement. All three of these prohibit the adoption of such measures. However, not unlike data flows provisions, they recognize the rights of the parties to adopt or maintain infringing measures necessary to achieve legitimate public policy objectives, provided that they are not applied in a manner that would constitute a means of arbitrary or unjustifiable discrimination or a disguised restriction on trade.

Conclusions Digital technologies discussed in this chapter have resulted in a shift in services trade from physical to virtual forms of supply that are enabling knowledge- and information-intensive services to be traded cross-border. They also have important implications for the manufacturing sector, particularly, but not only, IoT and 3D printing technologies. Sophisticated data analytics has already demonstrated the capability to enhance knowledge of customers, their preferences and purchasing habits in ways that are only beginning to bear potential for transforming both goods and services trade. Albert Einstein is often quoted as saying, ‘information is not knowledge’, but, rather, it becomes knowledge with imagination, analysis, innovation and experience. A corollary is that raw data is not information until it has been processed and transformed in an intelligible and useful form; that is, data itself is not knowledge, but rather is transformed into knowledge. Digitization makes this process easier and more rapid. Over-reactions to these transformations hold the risk of disproportionate policy responses which may harm not only trade, but also the

      



associated transmission of knowledge across borders. In many respects, the most salient features of the technologies behind today’s digitization, and its continued potential, are less a uniqueness of technologies themselves, but rather the effect of greatly enhanced potential for trade. In what ways does this transformation require new or enhanced rules of trade and engagement? Whatever trade rules may be employed, existing or new, one salient point is that strengthened and enhanced efforts at cross-border collaboration among governments may be needed to complement national initiatives.

Bibliography International Telecommunications Union (ITU), 2017 “ICT Facts and Figures 2017”, available online at https://www.itu.int/en/ITU-D/Statistics/Pages/ facts/default.aspx. Roy, Martin. (2017), “The Contribution of Services Trade Policies to Connectivity in The Context of Aid for Trade”, WTO Staff Working Paper, ERSD-2017-12. Tuthill, Lee (2016), “Cross-border Data Flows: What Role for Trade Rules?”, in Sauvé and Roy (eds), Research Handbook on Trade in Services, Edward Elgar, Cheltenham, UK. United Nations et al. (2010), Manual on Statistics of International Trade in Services 2010. WIPO, World Intellectual Property Report 2017: Intangible Capital in Global Value Chains, Geneva, 2017. WTO, Appellate Body Report, China – Measures Affecting Trading Rights and Distribution Services for Certain Publications and Audiovisual Entertainment Products, WT/DS363/AB/R, adopted 19 January 2010, DSR 2010:I. WTO, Appellate Body Report, United States – Measures Affecting the CrossBorder Supply of Gambling and Betting Services, WT/DS285/AB/R, adopted 20 April 2005, DSR 2005:XII, p. 5663 (and Corr.1, DSR 2006:XII, p. 5475). WTO, Panel Report, China – Certain Measures Affecting Electronic Payment Services, WT/DS413/R and Add.1, adopted 31 August 2012, DSR 2012:X, p. 5305. WTO, Panel Report, China – Measures Affecting Trading Rights and Distribution Services for Certain Publications and Audiovisual Entertainment Products, WT/DS363/R and Corr.1, adopted 19 January 2010, as modified by Appellate Body Report WT/DS363/AB/R, DSR 2010:II, p. 261. WTO, Panel report, Mexico – Measures Affecting Telecommunications Services, WT/DS204/R, adopted 1 June 2004, DSR 2004:IV, p. 1537.

   ,       WTO, Panel Report, United States – Measures Affecting the Cross-Border Supply of Gambling and Betting Services, WT/DS285/R, adopted 20 April 2005, as modified by Appellate Body Report WT/DS285/AB/R, DSR 2005:XII, p. 5797. WTO, Reference Paper on Basic Telecommunications (WTO Job document 2104, dated 24 April 1996).

A3.1 Specific commitments across selected sectors, by modes of supply number of schedules with specific commitments (out of 152)



Mode 1



Business services Professional services Legal services Accounting/auditing/ bookkeeping services Engineering services Computer services Consultancy services related to the installation of computer hardware Software implementation services Data processing services Database services Other R&D services R&D services on natural sciences

Mode 2

Mode 3

Mode 4

Full

Partial

Unbound

Full

Partial

Unbound

Full

Partial

Unbound

Full/ Partial

Unbound

77 88

25% 42%

66% 40%

9% 18%

35% 55%

60% 36%

5% 9%

13% 27%

82% 72%

5% 1%

97% 92%

4% 8%

89

55%

30%

15%

71%

17%

12%

48%

49%

2%

97%

3%

84

79%

7%

14%

85%

6%

10%

79%

20%

1%

98%

2%

89

74%

13%

12%

81%

12%

7%

71%

28%

1%

97%

3%

87

72%

15%

13%

80%

13%

7%

70%

29%

1%

97%

3%

80 62

78% 71%

9% 26%

14% 3%

86% 74%

8% 23%

6% 3%

75% 63%

24% 37%

1% 0%

97% 98%

3% 2%

42

71%

17%

12%

69%

19%

12%

64%

36%

0%

95%

5%



Other business services Advertising services 70 Management 83 consulting Communication services Postal services 14 Courier services 65 Telecommunication services Voice telephone 100 services Packet-switched data 93 transmission services Circuit-switched data 94 transmission services Private leased circuit 89 services Online information 85 and database retrieval Online information 70 and/or data processing

61% 80%

29% 11%

10% 10%

64% 84%

29% 10%

7% 6%

50% 80%

49% 19%

1% 1%

100% 100%

0% 0%

36% 54%

57% 34%

7% 12%

57% 60%

36% 34%

7% 6%

21% 43%

71% 57%

7% 0%

79% 95%

21% 5%

24%

70%

6%

51%

44%

5%

18%

81%

1%

95%

5%

34%

62%

4%

55%

41%

3%

21%

78%

1%

96%

4%

35%

60%

5%

56%

40%

4%

22%

76%

2%

96%

4%

36%

61%

3%

58%

39%

2%

24%

75%

1%

97%

3%

50%

45%

5%

63%

29%

8%

37%

60%

2%

98%

2%

55%

38%

7%

69%

23%

8%

42%

56%

1%

97%

3%

Source: WTO analysis based on GATS Schedules – available at i-tip.wto.org/services.

4 Intellectual Property and Digital Trade – Mapping International Regulatory Responses to Emerging Issues     .      -       Abstract This chapter explores how regulatory responses to emerging IP issues in digital trade may develop at the international level and in particular how existing mechanisms might influence the chances of developing internationally agreed rules in this regard. The primacy of state sovereignty in intellectual property up to the late 19th century gave way to the important WIPO treaties, which still retained some independence of member states and based international regulatory responses directly on national experience. While more regulatory sovereignty was ceded in TRIPS, the WIPO Copyright Treaty and the WIPO Performances and Phonograms Treaty, the adoption of non-binding instruments (such as the WIPO Joint Recommendations in the area of trademarks) show the limits of decision making by consensus. International non-state solutions such as the Uniform Domain-Name Dispute-Resolution Policy (UDRP) established by the Internet Corporation for Assigned Names and Numbers (ICANN) have introduced separate, technically determined solutions to specific IP issues. Proliferating free-trade agreements (FTAs) have emerged as a new platform to agree to IP-related regulatory responses that can be used to project the national solutions of a few dominant FTA-partners. However, these FTAs have also served to give legally binding status to internationally agreed non-binding recommendations. These diverse approaches are apparent in recent IP-regulatory responses to emerging digital issues that are particularly relevant for digital business models, including inter alia Internet service provider (ISP) liability, “safe harbour” provisions and the issue of orphan works, where there appears to be less agreement. Still further behind to reaching any kind of agreement are the emerging issues of online exhaustion, data mining and IP-related questions of artificial intelligence.

Introduction The genesis of international regulation in any subject area is complex and multifaceted. Whether a particular rule or practice succeeds in 

        



garnering international agreement is influenced by varying constellations of interests, stakeholders and institutions, and the fragile interaction between national jurisdictions and regional or international agreements. Internationally agreed regulation is particularly advanced in the area of intellectual property (IP), where detailed treaties set binding international standards on subject matter, scope of protection and enforcement of most types of IP rights. International registration systems exercise a harmonizing influence on procedures and formalities. The intangible nature of IP meant that it was one of the first regulatory areas to react to the challenges posed by digital networks and their consequences for territoriality and the reproduction of digital works. While many of the principles underlying existing IP regulation proved sufficiently adaptable to the digital world, many IP-related issues that have emerged since the advent of the Internet and digitization are triggering new or adapted regulatory responses at various levels. Some matters have already resulted in international agreements or uniform multilateral practices. Other – more recent – issues are still receiving different treatment in different jurisdictions and international consensus on how to regulate them is not (yet) in sight. To understand how international regulatory responses to IP-related challenges in the area of digital trade might develop at the international level, and what factors influence the locus and nature of such responses, this chapter first describes the rule-making processes that have shaped the international IP regime since its origins in the late nineteenth century, including recent examples of tools and legal instruments that address digital challenges. The chapter then reviews the status of selected current digital issues, including ISP liability and online exhaustion, from the perspective of mechanisms that might lead to internationally agreed responses.

The Changing Pattern of International Rule-Making in Intellectual Property The proverbial lag of regulatory and policy responses to technological developments is a frequently evoked stereotype that has become particularly popular – and true – in the area of digital technologies, where the rate and profundity of change has reached breakneck speed. This ‘tech lag’ phenomenon merely describes the time delay that national institutions face in identifying, formulating, issuing and implementing relevant policy or regulatory reactions to technological developments.



 .     -                   

However, even further complexity is usually associated with international regulatory responses, which – while having the advantage of disproportionate regulatory synergy across different markets – encounter myriad additional challenges as to legitimate fora, compatibility of legal systems, enforcement and adjudication of disputes.

The Early Ideal of Universal Codification and the Primacy of Sovereign Treatymaking The WIPO Paris and Berne Conventions, together with the TRIPS Agreement, form the foundations of today’s international IP system. Their substantive and institutional development largely shaped the advancement of internationally agreed IP rules during the twentieth century.

The Berne and Paris Conventions as the Foundation of International IP Regime The Berne and Paris Conventions resulted from the intangible nature – and hence easy international mobility – of inventions and literary works. The popularity of ‘Universal Exhibitions’ in the second half of the nineteenth century, and the corresponding surge of international exchange of industrial inventions and literature, soon united the relevant actors – industrialists and authors/publishers1 – to call for ‘universal’ unified regulation of authors’ and inventors’ rights. The resulting intergovernmental organizations (the Berne and Paris Unions) and conventions remained the prime locus for formulating international IP rules for the next century. While the early ideal of truly universal codification2 soon gave way to more realistic – and moderate – objectives, the early spirit of finding 1

2

For the role of the Association Artistique et Littéraire (ALAI) in the initiation of the Berne Union, see Ricketson and Ginsburg, (2006), pp. 49–58. See the programme of the Vienna Patent Congress 1873, Appendix A, Webster’s Report, page 400 (cited in Ricketson (2015), p. 34): ‘All the views, however, even those of the partisans of patent protection, unite invariably and unexceptionably in this, that the protection of the rights of inventors needs new forms corresponding to the altered international commercial relations; and the solution of this question of reform should not be aimed separately as hitherto, by each state of the great commercial area, but rather, that a complete solution common to all states should be accomplished by international agreement.’; regarding the Berne Convention, see the debate of the German motion in favour of a universal codification at Actes 1884, pp. 28–29, cited in Ricketson and Ginsburg (2006), p. 138.

       



common rules through the exchange of technical experts directly at the international level remained reflected in the institutions’ structure and operations.3 Exemplifying a form of international regulatory activity, these conventions – and the institutions administering them – provided a unique intergovernmental forum for technical experts in the relatively well-defined areas of IP to discuss and strive to formulate common regulatory solutions directly at the international level, which – if successful – became international treaty law. Participants in these deliberations were mostly like-minded national IP experts that subscribed to the common goal of international harmonization around appropriate subject matter principles and solutions. The iterative process of successive treaty revisions of the Paris and the Berne Conventions, which remained the focus of efforts up until the second half of the twentieth century, further reflects the lingering spirit that a ‘universal regulation’ of the entire area of industrial property or copyright could be contained in a single body of rules agreed among sovereign states. Political and policy considerations did influence divergent national positions. Yet in expert deliberations, and certainly in the dynamics of revision conferences, it is clear that the limitation of the treaty scope to the relatively confined area of IP and the traditional influence of practical technical questions (rooted in the early influence of authors, publishers and inventors) meant that the influence of strategizing and external political considerations could remain limited. The historical context of these treaties is complex and cannot be viewed in a simple linear pattern. Yet these arguably dominant traits of the early stages of international IP treatymaking help explain the surprisingly early and dynamic genesis of these conventions, which represent – then, as today – a degree and depth of international codification unparalleled in any other area of private law. The early and successful establishment of the Berne and Paris Unions with their expert-driven and unified regulatory approach made this model of sovereign treatymaking the natural focus for international IP regulation for most of the twentieth century, next to which alternative institutions – or even alternative approaches to treatymaking in this area – remained limited in scope.4

3

4

The unions were organized in ‘groups of experts’ that formed the precursors to today’s WIPO Standing Committees. This is not to diminish alternative efforts such as the Universal Copyright Convention (UCC), that pursued equally ‘universal’ ambitions.



 .     -                   

By the time digitization and the Internet emerged in the 1990s, however, this traditional model of consensus-based treaty revisions based on expert opinion had reached certain limits. Concerns of developing countries had dominated the revision conferences in the 1960s and 1970s, and the iterative revisions of entire conventions had become difficult – and, while still hoped for in the 1990s, effectively came to a halt after the 1971 revision of the Berne Convention.5 The emerging and accelerating technological changes, and the corresponding new means of exploitation, now led to more divergent national solutions and approaches that could no longer be easily bridged by national expert discussions at the international level, as had been the case in the early years. Hence the provisions of the 1971 Berne Convention remained far from the original ideal of ‘uniformity of protection’ with implications for coverage of emerging digital issues. Gaps persisted, in particular with regard to a right of distribution of copyright protected works. The formulation of a number of protected rights still permitted considerable variations in national interpretation.6 Given the wide discretion left to countries with regard to copyright exceptions, and the lack of clarity as to whether software fell within the scope of the Convention, much of the effective protection of authors was effectively left to the application of national treatment. While member countries maintained the characteristically expertdriven approach in the context of WIPO by establishing committees of government experts on a number of salient issues – such as computer software as early as the 1970s and 1980s – these efforts did not bear fruit until after the conclusion of the TRIPS negotiations, when some of the work from the ‘guided development’7 period served as a basis of the 1996 WIPO Internet Treaties.

TRIPS: A Continuation of Unified ExpertDriven Treatymaking Although the inclusion of IP subject matter in the multilateral negotiations of the GATT Uruguay Round represented a significant shift from the historical pattern of treatymaking, the methodology and the outcome of the negotiations of the TRIPS Agreement still resemble the model of 5

6

7

See the account of ‘uneasy truce’ that no one wanted to reopen by beginning new debates, in Ricketson and Ginsberg (2006), p. 140. This was the case in the areas of reproduction, broadcasting and cable distribution. See Ricketson and Ginsberg (2006), p. 138. Ricketson and Ginsberg (2006), p. 143.

       



unified expert-driven treatymaking. The factors that facilitated the shift of focus from WIPO to the GATT/WTO have been amply commented upon.8 From the perspective of international IP regulatory activity, these factors fall into two categories: those related to the limits of the traditional locus and methodology of international IP regulation in WIPO; and those relating to the systemic dynamics of multilateral trade negotiations. For WIPO, the lack of progress since the last revisions of the Paris and Berne Conventions in 1967 and 1971 respectively meant that there was little prospect that remaining gaps in protection, notably on enforcement and dispute settlement, could be closed in the traditional fashion in the foreseeable future. The protracted experience of the last revision conferences, and the failure of attempts to conclude ‘special agreements’ in the 1980s – notably on software9 – opened some Member countries’ minds to alternative approaches of treatymaking, although much convincing remained to be done before agreement on negotiating a comprehensive IP instrument in GATT was reached in the course of the Uruguay Round.10 In the trade negotiations area, the characteristic trade-off dynamic of the Uruguay Round negotiations meant that, in contrast to the WIPO context, non-IP considerations were able to exercise more considerable – and more direct – leverage on countries’ willingness to discuss IP. At the time, this concerned first and foremost the question of whether a fullyfledged TRIPS outcome should be part of the Uruguay Round at all. The mandate to negotiate comprehensive ‘standards and principles concerning the availability, scope and use of trade-related IPRs’ was agreed in April 198911 – after considerable hesitation on the part of a number of both developed and developing countries – only when the prospective benefits of the Uruguay Round in areas such as textiles, agriculture and tariffs became clearer, and a refusal to negotiate on IP would have endangered the benefits of the overall outcome.12 It is less evident that this trade-off leverage had any effect on individual substantive IP questions in the same way as is argued in the context of recent bilateral treaties.13 Rather, it seems that negotiating countries on the whole saw IP negotiations as a whole as a trade-off for lucrative agricultural market 8 9 10 11

12 13

Ibid., p. 155, note 44. Ibid., p. 144. See Otten (2015). GATT Document MTN/TNC/11, Uruguay Round – Trade Negotiations Committee – Mid-Term Meeting, 21 April 1989. Otten (2015), pp. 62 and 74. See ‘Regional Trade Agreements: Regulatory Responses to Digital Challenges’ below.



 .     -                   

access offers, hence requiring progress on the latter before continuing the former during the successive negotiation sessions.14 Negotiations on individual IP issues remained largely guided by the – offensive or defensive – substantive interests, patiently explained by IP experts, of the negotiating countries who lined up in groups – often defying the traditional north–south narrative – of common positions around copyright, enforcement or geographical indications, to name just a few.15 From the perspective of international IP regulation, therefore, and despite this new context of international IP regulation in the Uruguay Round, the substantive outcome on IP within the TRIPS Agreement followed in many ways the characteristics of the treatymaking process under the WIPO Conventions. Although the negotiations moved much faster and in a less formal manner, and took place outside the WIPO ecosystem of government expert committees and working groups, national IP experts played a central role in the framing and the conduct of TRIPS negotiations. Furthermore the historic ideal underlying the conventions, namely of ‘universal codification’ seems to have been even more successful in the unifying edifice of the TRIPS Agreement. Incorporating tel quel not only most of the provisions of the Berne and Paris Conventions, but also parts of the Rome Convention and the Washington Treaty, and supplying significant substantive additions and clarifications, the TRIPS Agreement remains the single most comprehensive international treaty on IP today – covering substantive protection and enforcement of most conventional categories of IP in a single treaty. Although negotiated just as early digitization issues were arising,16 the TRIPS Agreement contains few specific regulatory responses in this area. Notable is the first international confirmation that software is covered by copyright, resolving an issue discussed in WIPO in the 1980s. The specific formulation ‘whether in source or object code’ clarified that protection extended also to the machine-readable binary code in which software was usually distributed – which some national jurisprudence had excluded,17 hence rendering protection ineffective. By stipulating protection ‘as literary works’, Article 10.1 TRIPS ensured that computer

14 15

16

17

See the account of the Brussels breakdown of negotiations in Otten (2015), p. 67. See the country experiences reported in Watal and Taubman (2015) The Making of the TRIPS Agreement (WTO 2015). Although signed only in 1994, the bulk of the TRIPS text was negotiated already in the 1980s. See Gervais (2012). See Blomqvist (2014), p. 88.

       



programs – with their arguably functional character – would not be categorized as works of applied art, which would have permitted a reduced minimum term of protection,18 or even exclusion from protection altogether – albeit while risking material reciprocity – under the Berne Convention (1971).19 TRIPS Article 10.2 on ‘compilations of data and other material’ – while deciding against requiring sui generis database protection at the international level20 – specifically acknowledges that compilations of data or other material ‘whether in machine-readable or other form’ are to be protected if they constitute intellectual creations by reason of selection or arrangement, even if the data or the material itself is not copyright protected. Against this – admittedly simplistic and selective – account charted above it would seem that, despite the institutional shift from a body exclusively dedicated to IP into the multilateral trade regime, the TRIPS negotiations represent nevertheless the culmination of the traditional model of international IP regulation in which (i) largely expert-driven negotiations directly at the international level develop rules in pursuit of – ideally – (ii) a unified and universal body of rules in the shape of conventions or comprehensive texts, that are then agreed to as (iii) international treaties by sovereign states. The limits of this model arguably showed already in the earlier slowdown after the conventions’ last revisions and it could be argued that it was only the context of the tradeoffs available in the Uruguay Round that achieved the renewed momentum. The TRIPS outcome nevertheless shows the continued salience and attractiveness of the ‘universal sovereign treaty model’ in the early 1990s, although other – more limited – selective subject matter agreements would also have been possible, and were in fact contemplated.21

The Emergence of New Modes of Regulatory Responses The culmination of the model described above coincided with the early development of the World Wide Web, and therefore did not address the new issues raised by widespread use of the Internet, particularly the 18 19

20 21

Article 7(4) Berne Convention (1971). Article 2(7) of the Berne Convention (1971); see also Blomqvist (2014). For an overview of the negotiating dynamics on copyright protection of software at the time of the Uruguay Round, see Wager (2015). See the background in Gervais (2012). See the GATT 1982 and 1986 Ministerial Declarations with a mandate for a code on counterfeit goods only.



 .     -                   

online distribution of protected works. In responding to these challenges, the traditional model began to give way to more varied regulatory responses.

The WIPO Internet Treaties: An Issue-Driven International Solution Although work on the challenges presented by new technological developments had begun in WIPO already in the 1970s, progress had been slow, and momentum was lacking particularly once the Uruguay Round mandate to negotiate a comprehensive IP agreement had taken shape in late 1989. As TRIPS did not address many of the new technology issues in detail, momentum returned to WIPO where an astonishing spurt of concerted effort produced the WIPO ‘Internet Treaties’, namely the WIPO Copyright Treaty (WCT) and the WIPO Performances and Phonograms Treaty (WPPT), less than two years after the Uruguay Round concluded.22 The Internet Treaties took significant steps towards adapting international rules for the protection of copyright and the rights of performers and producers of sound recordings to the digital revolution, and in particular to the distribution of copyright material over the Internet.23 These treaties establish an international agreement24 that storage of works in a digital form in an electronic medium constitutes a reproduction. Despite alternative approaches discussed in the run-up to the Diplomatic Conference, signatories ultimately agreed that the respective reproduction rights, of authors, performers and of phonogram producers – as well as the permitted exceptions – fully apply to the digital environment.25 While this ensured that digital exploitation and storage was covered by the traditional reproduction concept, the specific mention of the applicability of permitted exceptions, and a certain leeway in interpreting the term ‘storage’, was understood to permit justified exceptions regarding transient or incidental reproductions,26 as may occur in 22

23 24 25

26

The WCT and the WPPT were adopted on 20 December 1996 and entered into force on 6 March 2002 and 20 May 2002, respectively. Wager (2015). By March 2020, the Internet Treaties had been signed by 104 countries. For copyright, see Article 1(4) WCT and agreed statement. For performers’ and phonogram producers’ rights, see Articles 7 and 11 of the WPPT that confirm and expand the reproduction rights provided for in Article 7.1 (for performers) and Article 10 (for phonogram producers) of the Rome Convention, and the accompanying agreed statement. See WIPO Handbook (2004), 5.220 and 5.563.

       



the operation of digital networks.27 An agreed statement to the TRIPSstyle three-step test that was now introduced to cover exceptions of the Internet Treaties and the Berne Convention28 explicitly highlighted the objective to permit the development of new exceptions and limitations appropriate for the digital network environment.29 The treaties further enshrine the central principle that the transmission of works in digital networks is the object of an exclusive right of the author or other rights owner. In formulating this ‘right of making available’ in a manner that captured the technological specificity of Internet distribution, yet remained neutral as to which traditional rights were the basis of this new obligation, the treaties bridged considerable diversity that existed in national jurisdictions as regards the recognition and application of a distribution right or a right of communication to the public.30 As the Berne Convention (1971) coverage of both these rights was limited,31 the Internet Treaties further endeavoured to fill these gaps by explicitly extending the right of communication to the public to cover the making available of works in digital networks (Article 8 WCT, Articles 10 and 14 WPPT), and by explicitly stipulating a distribution right (Article 6 WCT, Articles 8 and 12 WPPT), thus clarifying the hitherto heterogenous interpretation of the international standard. This so-called ‘umbrella solution’32 – to formulate a new making available right in a ‘neutral, legal characterization-free description’33 – ensured 27

28

29

30 31

32

33

This permitted, in principle, the same treatment as for ‘ephemeral fixations’ by broadcasting organizations under Article 15.1 (c) of the Rome Convention (1961). It is Article 10(2) WCT that extends the test to govern exceptions and limitations to all Berne Convention rights. The agreed statement to Article 10 WCT – also covering Article 16(2) WPPT – reads: ‘It is understood that the provisions of Article 10 permit Contracting Parties to carry forward and appropriately extend into the digital environment limitations and exceptions in their national laws which have been considered acceptable under the Berne Convention. Similarly, these provisions should be understood to permit Contracting Parties to devise new exceptions and limitations that are appropriate in the digital network environment. It is also understood that Article 10(2) neither reduces nor extends the scope of applicability of the limitations and exceptions permitted by the Berne Convention.’ See WIPO Handbook (2004). In the Berne Convention (1971) a distribution right is only explicitly granted with respect to cinematographic works (Article 14(1)(i)), although some jurisdictions considered such a right a necessary – and thus implicit – corollary of the right of reproduction. The right of communication to the public only extends to certain forms of communication of works (Article 11bis(1)). The term was reportedly coined by Mihály Ficsor, Secretary of the Diplomatic Conference. See Schlesinger (2010), p. 180. Mihály Ficsor (2002), § C.8.06.



 .     -                   

that both possible bases for such an obligation – the right of communication to the public and the distribution right – were now equally enshrined in the international standard of copyright protection. This artful solution gave national legislatures considerable flexibility in characterizing the exclusive rights involved.34 While Article 10 WCT associates the new right of making available with the right of communication to the public – and many countries have chosen to implement that way35 – the negotiation history36 and subsequent commentary37 clarifies that the ‘umbrella solution’ equally permits signatories to implement the making available right through applying the exclusive distribution right to electronic transmissions of works.38 Indeed, in that regard, the United States observed that: a ‘distribution’ approach more closely approximates the real market impact of on-demand and other online transactions, in which the recipient of the transmission often ends up with a copy of a work that he or she did not have before.39

The United States therefore saw no need for legislative action to implement the new right, since ‘its broad reading of the distribution right, in conjunction with the reproduction right [. . .] would cover the act of “making available” as that is defined in the WCT (and the WPPT)’.40 Some of the challenges that would be raised by the implementation of these new rights, and which do indeed occupy courts and law-makers today – such as ISP liability and the question of online exhaustion – were anticipated in some of the agreed statements adopted with the Internet Treaties. The newly confirmed right of distribution was accompanied by a statement seeking to clarify that the ‘copies’ subject to the distribution (and

34 35

36 37 38 39 40

See Schlesinger (2010). Ibid. See also an early Survey on Implementation Provisions of the WCT and the WPPT, document SCCR/9/6, 25 April 2003, which mentions that the right of making available had been implemented as a right of communication in 19 out of 39 laws reviewed, and ‘under or in conjunction with’ the right of distribution in only five out of 39 laws reviewed. WIPO Handbook (2004), 5.226. Schlesinger (2010) and von Lewinski (2015), § 8.10.10. Schlesinger (2010), p. 181. For the WPPT, see von Lewinski (2015), § 8.10.10. Schlesinger (2010). Ibid. See also US Copyright Office (2016), pp. 74–75 confirming the sufficiency of current statute: ‘In general, where a party offers members of the public access to a work in the form of a download, the offer implicates the right of distribution.’ ‘. . . Congress understood such conduct to be an infringement of the distribution right.’

       



rental) right were tangible objects.41 Together with the traditional understanding, expressed in the notes of the Basic Proposal, that ‘no rights are exhausted in connection with communication’ and that ‘exhaustion of rights is only associated with the distribution of tangible copies’,42 this provides the dogmatic background for asserting that in countries that consider ‘making available’ a communication to the public, electronic copies made available in digital networks would not be subject to exhaustion. However, the basic tenet of the ‘umbrella solution’ – namely that, despite the wording of Article 8 WCT, countries are free to implement the new exclusive making available right by extending the distribution right to cover electronic transmissions – leaves the dogmatic context more ambiguous for those who, like the United States, consider online downloads an act of distribution. This tension is recognized in some characterizations of the agreed statement, which seems to imply that countries’ freedom of choice regarding the implementation of the new right includes the freedom to have exhaustion apply to electronic copies ‘distributed’ through the making available right. The question may emerge whether this Agreed Statement [concerning Articles 6 and 7] conflicts with the ‘umbrella solution’ for transmissions in interactive digital networks, and, particularly, whether or not it excludes application of the right of distribution to such transmissions. The answer to this question is obviously negative. The Agreed Statement determines only the minimum scope of application of the right of distribution; it does not create any obstacle for Contracting States to exceed that minimum.43

As regards the issue of ISP liability, the new making available right in Article 8 WCT was accompanied by an agreed statement clarifying that the provision of physical infrastructure did not in itself amount to an act

41

42

43

‘Concerning Articles 6 and 7: As used in these Articles, the expressions “copies” and “original and copies”, being subject to the right of distribution and the right of rental under the said Articles, refer exclusively to fixed copies that can be put into circulation as tangible objects.’ See the Basic Proposal [. . .] to be Considered by the Diplomatic Conference (CRNR/DC/ 4 of 30 August 1996), para. 10.20: ‘It should be pointed out that no rights are exhausted in connection with communication to the public. Should communication of a work result in the reproduction of a copy at the recipient end, the work may not be communicated further to the public or distributed to the public without authorization. Exhaustion of rights is only associated with the distribution of tangible copies.’ Mihály Ficsor (2005), p. 11.



 .     -                   

of communication within the meaning of the treaty,44 which is intended to clarify the issue of the liability of service and access providers in digital networks.45 However, other than stating the rather obvious fact that providing infrastructure is not communicating to the public, this agreed statement is not determinative of who is liable for infringements of the exclusive right of making available stipulated in Article 8 WCT. This brief – and selective46 – overview of some of the Internet Treaties’ achievements in crafting creative regulatory responses to the challenges the digital environment posed for copyright and related rights highlights a certain departure from the traditional treatymaking pattern identified earlier. The negotiators did not pursue international regulation of copyright or related rights in a ‘unified and universal body of rules’, i.e. by pursuing another revision of the traditional conventions, but deliberately chose separate instruments (making use of the Berne Article 20 provision for special agreements), with their own treaty administration,47 that could be acceded to individually.48 Although this is certainly not to suggest an abandonment of consistency with the existing body of rules,49 this formal separation of instruments permitted a more focused – and arguably more ambitious – regulatory response to the specific challenges posed by new communication technology, and thus avoided the concerns associated with reopening the Berne Convention. The Internet Treaties, like the Uruguay Round conclusion a few years earlier, also coincided with a geopolitical situation more favourable to market opening, multilateral cooperation and to embracing technological 44

45 46

47 48

49

See agreed statement concerning Article 8, ‘It is understood that the mere provision of physical facilities for enabling or making a communication does not in itself amount to communication within the meaning of this Treaty or the Berne Convention. It is further understood that nothing in Article 8 precludes a Contracting Party from applying Article 11bis(2)’. WIPO Handbook (2004), 5.567. For a full overview of the regulatory breadth of the Internet Treaties, see the respective sections of the WIPO Handbook (2004). See Article 15 WCT and Article 24 WPPT for arrangements of the respective assemblies. Note that membership of the Berne or Rome Conventions was not a condition for eligibility to join the Internet Treaties (see Article 17 WCT and Article 26 WPPT). In March 2020, the WCT and WPPT were in force in 97 out of a total of 104 contracting parties, while the 1971 Paris Act of the Berne Convention was in force in 188 contracting parties. Indeed, the elaborate provisions on relationships with other treaties emphasize the link with the Berne and Rome Conventions, respectively, and ensure that the Internet Treaties may not be interpreted in a manner that is inconsistent with their individual preceding conventions. See Article 1 WCT and Article 1 WPPT.

       



change, which influenced countries’ willingness to agree on complex legal issues in an international forum. Many former communist or socialist economies were actively studying and seeking to establish societies based on market economy principles, which sometimes served as an inspiration for traditional developing countries to pursue similar ambitions.50 The conclusion of the TRIPS Agreement, which reconfirmed, unified and further built on the Berne and Paris Conventions, may thus have helped overcome regulatory stagnation in the field of IP, and represented a new platform from which new solutions for the challenges posed by the accelerating technological developments could be developed in the Internet Treaties. In some areas, countries were prepared to agree solutions that had not yet been legislated or contemplated in the national jurisdictions, and thus again represent expert-driven, substantive international law-making, rather than agreement on a lowest common denominator of existing national solutions.51 From a regulatory point of view, the issue-driven Internet Treaties thus represent a particularly successful instance of treatymaking that was able to benefit from thorough expert preparation in substance and favourable geopolitical circumstances to achieve consensus on a rapid and relatively comprehensive response to the copyright challenges of new communication technologies. While countries were able to build on this success to some extent with the conclusion of ‘the 3rd WIPO Internet Treaty’52 – the Beijing Treaty on Audiovisual Performances in 2012 – achieving treaty-level consensus on regulatory responses to digital challenges soon became more difficult and gave rise to the use of non-binding regulatory instruments.

The Emergence Of Non-Binding Instruments – the Joint Recommendations While work on the Internet Treaties was successfully concluded, alignment of interests in other areas was more elusive. Increased commercial Internet 50 51

52

Otten (2015), Wager (2015). The obligations on technical protection measures and digital rights management information (Articles 11–12 WCT and Articles 18–19 WPPT) have not been treated here. For an example of discussions in the context of subsequent implementation of these obligations see Geist (2010), and Ficsor (2009). See Mihály Ficsor, ‘The WIPO Internet Treaties and Copyright in the “Cloud”’, paper presented at the ALAI Congress in Kyoto, 2012.



 .     -                   

use and the emergence of new trademark laws and registration authorities, among other factors caused by technological and commercial developments, had brought into sharp focus the need for clarity on the practical definition of ‘well-known’ marks whose scope of protection in Article 6bis Paris Convention had just been extended by Article 16 of the TRIPS Agreement.53 Based on a mandate to study ‘all questions of relevance to the correct application of Article 6bis’ from the 1996–1997 WIPO Programme54 a Committee of Experts on Well-Known Marks – later the Standing Committee on Trademarks, Industrial Designs and Geographical Indications (SCT) – developed draft provisions. Up until its third session in 1997 the Committee of Experts foresaw the adoption of these draft provisions either through an international instrument or a recommendation by the WIPO General Assembly, ‘once sufficient agreement had been reached on such conclusions’.55 However, at around the same time WIPO members officially embraced the need to adapt to the pace of change in the field of industrial property by considering new options – i.e. other than treaties – for agreeing international common principles: Given the practical imperative for accelerated development and implementation of certain international harmonized common principles and rules in industrial property law, the future strategy for this main program includes consideration of ways to complement the treaty-based approach [. . .] If Member States judge it to be in their interests so to proceed, a more flexible approach may be taken towards the harmonization of industrial property principles and rules, and coordination of

53

54 55

See Kur (2013) remarking that ‘It was not uncommon during that phase that fortuneseekers succeeded in registering famous marks like “Dior” or “Cartier” and others, in order to extract money from the true proprietors when they tried to get a foothold on the same market. This of course created concern among the trademark community in the Western world, leading to warnings directed at the political instances in those countries that they were in violation of their obligations under Article 6bis Paris Convention and of Article 16 TRIPS, threatening to compromise their access to WTO membership. Trying to achieve what was expected of them most governments were keenly interested in obtaining specific and secure guidelines for checking which marks were to be considered as “well-known” in the meaning of the international agreements and hence must be excluded from trademark protection.’ WIPO Document AB/XXVI/2, Item 03(5). Memorandum on Protection of Well-Known Marks at the Third Session of the Committee of Experts on Well-Known Marks, 20–23 October 1997, WIPO Document WKM/CE/III/2: ‘Therefore, [. . .] it is left open whether those draft provisions would be adopted in the form of a recommendation of the said bodies or in the form of an international instrument such as a Protocol to the Trademark Law Treaty.’

       



administration, so that results can be achieved and applied more rapidly, ensuring earlier practical benefits for administrators and users of the industrial property system.56

Under this approach, WIPO member states agreed three non-binding recommendations57 between 1999 and 2001. In 1999, the Paris Union Assembly and the WIPO General Assembly adopted the Joint Recommendation Concerning Provisions on the Protection of WellKnown Marks (JR 1999), which provided criteria for the determination of whether a mark qualifies as ‘well-known’58 and established remedies for conflicts between well-known marks and other marks, business identifiers and domain names.59 Its relevance for digital trade arises from its recommended prohibition on requirements for use or registration of the mark in the member state as a condition for it to be considered well known60 there, while recommending the recognition of use and promotion or advertising – including on the Internet – as factors counting in favour of the well-known status of a mark.61 Thus the JR 1999 significantly shores up the right owners’ position in the digital marketplace, since advertising and use of the mark on the Internet is privileged vis-àvis actual local use in the relevant member state.62 The newly created SCT had also started more general work on ‘Trademarks and the Internet’ at its second session in 1999 which developed into the Joint Recommendation Concerning Provisions on the Protection of Marks, and Other Industrial Property Rights in Signs, on the Internet (JR 2001) adopted in 2001. These recommendations attempted to harmonize the interpretation and meaning of ‘trademark use’ that had remained undefined in the Paris Convention and the TRIPS Agreement, 56

57

58 59 60 61

62

WIPO Program and Budget for the biennium 1998–1999, document A/32/2-WO/BC/18/ 2, p. 86. Joint Recommendation Concerning Provisions on the Protection of Well-Known Marks (1999), a Joint Recommendation on Trademark Licenses (2000) and a Joint Recommendation Concerning the Protection of Marks, and Other Industrial Property Rights in Signs, on the Internet (2001). JR 1999, Article 2. JR 1999, Articles 4–6. JR 1999, Article 2(3)(i). JR 1999, Article 2(1)(b)(2) and (3). Note that ‘promotion’ was separately mentioned as a factor counting towards well-known status, also to overcome any ambiguity of whether promotion and advertising were regarded as ‘use’ of the mark. See the (explanatory – not ‘agreed’) Notes on Article 2, available at www.wipo.int/edocs/pubdocs/en/wipo_pub_ 833-accessible1.pdf. See Kur (2013).



 .     -                   

and had come into sharper focus with the contradiction between the principle of territoriality of rights and the global nature of the Internet. The present provisions are intended to be applied in the context of determining whether, under the applicable law of a Member State, use of a sign on the Internet has contributed to the acquisition, maintenance or infringement of a mark or other industrial property right in the sign, or whether such use constitutes an act of unfair competition, and in the context of determining remedies.63

In substance, the JR 2001 recommends resolving the question of whether Internet use of a sign can be considered as use in a certain territory on the basis of whether it constitutes a ‘commercial effect’ according to a list of factors.64 It also recommends a ‘notice and avoidance of conflict’ procedure for avoiding conflicts of right holders of identical or similar rights granted in different countries and their use over the Internet, under which right owners and other legitimate users are exempt from liability until they are notified of a conflicting right.65 It recommends that remedies for infringement be limited, as far as possible, to the territory in which the right is recognized, and only be available if the allegedly infringing use of the sign can be deemed to have taken place in that territory. In other words, the JR 2001 proposes that the ‘commercial effect’ of Internet use should serve as a yardstick for determining a ‘proportionate’ remedy.66 As international regulatory responses, the Joint Recommendations depart further from the traditional treatymaking model. They pursue distinct solutions to specific details within a single IP discipline, not a unified body of rules. And they are non-binding ‘soft law’, not binding international treaty instruments agreed to by sovereign states. The need for acceleration of rule-making to ensure ‘earlier practical benefits for administrators and users’67 that is cited in support of this new approach also seems to suggest that the traditional expert-driven process that would have been necessary to ensure consensus for binding treaty provisions may have given way to more interest-driven results68 that representatives could accept in the absence of binding commitments.

63 64 65 66 67 68

JR 2001, Preamble. Article 3 JR 2001. Article 9 JR 2001. WIPO Handbook (2004) 5.751. WIPO Program 1998–1999 (n. 56). Kur (2013) observes: ‘Driven not least by international stakeholders’ associations, the text is pointedly right-holder friendly, without offering much legal certainty in return.’

       



Their non-binding nature may have facilitated their adoption but has not impeded their influence on national IP policymaking. Indeed, the standing of these instruments as recommendations from two authoritative WIPO bodies has led to their direct incorporation into legislative projects in some countries,69 or to them being taken into account in national70 jurisprudence. More recently, their inclusion in IP chapters of Free Trade Agreements have turned these non-binding recommendations into binding and enforceable commitments of an increasing number of countries.71

A Technically-Determined Solution to Specific IP Issues: The UDRP The contemporaneous privatization of the hitherto government-run structure of the Domain Name System (DNS) and the creation of the Internet Corporation for Assigned Names and Numbers (ICANN) led to the creation of an entirely different IP-related regulatory response in the area of the Internet’s infrastructure. The use of domain names as ‘user friendly’ labels to connect to numerical Internet Protocol addresses on the global Internet inevitably led to conflicts with the existing territorial protection of trademarks under the national laws.72 The National Science Foundation (NSF) was authorized to permit commercial activity on the NSFNET – the nonmilitary part of the Internet it was administering – in 1992,73 and the ensuing interest of commercial actors led to a rapid expansion of the Internet.74 Network Solutions Inc. (NSI), a private company in charge of registering and administering the most valuable generic Top Level Domains (gTLDs)75 – including .com – on behalf of the NSF on a first come, first served basis, originally saw no substantive role for the 69

70

71 72 73

74

75

See for example the Trade Marks Act of the Republic of India, Chapter II, 11(6) to (10) referring to JR 1999. See section 9.3.3 The Long Arm of Non-Binding Decisions at WIPO, in Kwakwa and Talbott (2013). See below Regional Trade Agreements: Regulatory Responses to Digital Challenges. See generally Dinwoodie (2000). See Scientific and Advanced-Technology Act of 1992; Pub. L. 102-476 § 4(9), 106 Stat. 2297, 2300 (codified at 42 USC § 1862 (a)). For an overview of the Domain Name System and the evolution of its governance structure, see Petillion and Janssen (2017), chapter 2. NSI, a private American company, had received a directive from the InterNIC in 1993 to administrate the generic Top Level Domain Names (gTLDs): COM, EDU, GOV, INT, NET, ORG and MIL. See Haas (2009).



 .     -                   

registration authority in the resolution of such disputes.76 However, as disputes over the Internet’s naming system soared77 and criticism over the absence of competition in the registration process grew – and as NSI began charging user fees for domain name registration78 – NSI introduced a first Domain Name Dispute Policy in 1995. Based on the contractual relationship between the domain name applicant and the registrar, to which it was annexed, this policy sought to avoid liability and prevent potential lawsuits of third parties against NSI.79 NSI’s role as sole DNS registrar, and its Domain Name Dispute Policy, attracted sustained criticism over its monopoly position and, from the trademark community, its archaic dispute policy. The policy’s approach to resolving conflicts between trademarks and domain names on the basis of registration priority80 stood in stark tension with the longstanding US first-to-use tradition, by which rights in a mark are based on first use, not registration.81 While the Clinton administration’s ensuing decision to ‘privatize’ the Internet, and NSI’s subsequent loss of registrar monopoly with the creation of ICANN and a new governance structure, was taken in the much larger policy context of Internet governance,82 the tension between the ‘commercial’ trademark owner community and the ‘academic’ engineering community that had created the network was a visible part of this defining period for the Internet’s naming system.

76

77

78

79

80

81 82

Response to RFC (Request for Comments) 1561 from March 1994: ‘In case of a dispute between domain names registrants as to the rights to a particular name, the registration authority shall have no role or responsibility other than to provide the contact information to both parties. The registration of a domain name does not have any trademark status. It is up to the requester to be sure he is not violating anyone else’s trademark’, cited in Haas (2009). In a 1998 Wired.com article, Dave Graves, Network Solutions’ director of business affairs, is quoted: ‘Our dispute policy has been invoked about 1,700 times in the two years that it has been in existence’, he said, ‘Compared to a database of over 1.6 million domain names, that represents about one-tenth of 1 percent.’ See Stutz (1998). Petillion and Janssen (2017) Chapter 2, n 57. NSI was permitted to charge USD 50 per year per second-level domain. Haas (2009). See also INTA White Paper (1997) Section E. NSI Dispute Policy: What Happens When Domain Names and Trademarks Conflict. Haas (2009): ‘NSI determined the “creation date” of the domain name registration. If it was prior to the trademark, no action was taken.’ For a thorough conceptual discussion of this tension see Dinwoodie (2000), p. 515. For the larger context see the introductions to the US Green Paper (1998), and the US White Paper (1998).

       



The International Trademark Association’s (INTA) 1997 White Paper,83 in criticizing the state of affairs under the NSI policy, opined that At the threshold is the issue that the Internet was not created solely for commercial enterprise and that domain names should not be the exclusive province of trademark owners. Thus, those with legitimate non-trademark interests in second level domain names have to be accommodated along with the rights of trademark owners. Similarly, the interests of owners of trademarks which exist in commerce concurrently for non-related products or services must be considered as do those of famous marks. Finally, it is not necessary to have a second level domain name to do business effectively or successfully on the Internet, and thus while it may be preferable to acquire the second level domain name of your choice, it is not an absolute right and may have to bend to accommodate competing interests.84

INTA’s preferred solution was thus not to create a separate substantive domain name dispute resolution policy at all. Rather, a sui generis approach would permit a workable procedural approach to domain registration and disputes, while all substantive issues ‘will be left to the courts and trademark tribunals’ and their already growing body of domain name dispute jurisprudence,85 thus ensuring that traditional trademark law would apply to these conflicts. The NSI’s director of business affairs disagreed with trademark lawyers’ criticism of their revised Dispute Policy. ‘[T]here’s nothing at all in law that says having a trademark and the rights to that trademark also provide an automatic right to a domain name,’ said Graves. ‘There’s nothing that says that.’86

The alternative governance model known as gTLD-MoU – competing with the much-criticized87 US Department of Commerce’s Green Paper proposal of 30 January 199888 – built on previous criticism of NSI’s dispute policy. It focused on a self-regulatory and market-driven mechanism enforced by administrative challenge panels (ACP) that would reside within WIPO’s Arbitration and Mediation Center.89

83 84 85 86 87 88 89

INTA White Paper (1997). INTA White Paper (1997). INTA White Paper (1997). David Graves, NSI cited in Stutz (1998). Petllion and Janssen (2017), chapter 2, A.2.b.(3)(a). US Green Paper (1998). Petillion and Janssen (2017), chapter 2, A.2.b.(3)(a). See also Komaitis (2010), p. 75.



 .     -                   

While this MoU, as well as the envisaged governance association with the International Telecommunications Union (ITU), did not garner the required governmental support of the United States,90 many of the characteristics of its dispute resolution approach91 would be replicated in the Uniform Domain Name Dispute Resolution Policy (UDRP) that was developed under WIPO auspices and eventually adopted by ICANN under the arrangements of the US Department of Commerce’s White Paper of 5 June 1998,92 in which the Clinton administration’s privatization policy for the DNS was settled. The US White Paper’s solution to the ‘Trademark Dilemma’93 was to ask WIPO to initiate a ‘balanced and transparent’ process to develop a uniform approach to resolving trademark/domain name disputes involving cyberpiracy (not conflicts between competing trademarks), albeit with precise recommendations that sought to address problems and criticism identified under previous experiences. These included the maintenance of an up-to-date database of domain name owners’ contact information that should be accessible to trademark owners in order to identify and contact potential infringers, and development of process to exclude famous trademarks from being used as domain names except by the trademark holder. It further recommended that, at the time of registration or renewal, domain name registrants submit ‘infringing domain names’ to a court of law in a certain jurisdiction,94 while for cases ‘involving cyberpiracy and cybersquatting’ they would submit to and be bound by the alternative dispute resolution systems that the new corporation (i.e. ICANN) would adopt following the WIPO process.95

90 91

92 93 94

95

Ibid. See also US White Paper (1998). See Komaitis (2010): ‘The proposal conceived of a mechanism that would be hybrid in nature, would not replace traditional means of adjudication and would see administrative panels adjudicating disputes under certain objective standards and criteria.’ US White Paper (1998). US White Paper (1998), section 8. ‘1) Domain registrants pay registration fees at the time of registration or renewal and agree to submit infringing domain names to the authority of a court of law in the jurisdiction in which the registry, registry database, registrar, or the “A” root servers are located.’ US White Paper (1998) Recommendations. ‘2) Domain name registrants would agree, at the time of registration or renewal, that in cases involving cyberpiracy or cybersquatting (as opposed to conflicts between legitimate competing rights holders), they would submit to and be bound by alternative dispute resolution systems identified by the new corporation for the purpose of resolving those conflicts. Registries and Registrars should be required to abide by decisions of the ADR system.’ US White Paper (1998) Recommendations.

       



Finally, the rights of both domain name registrants and trademark owners under national laws should remain unaffected.96 After a brisk process of nine months, and taking into account reactions to an interim report,97 WIPO delivered its final report98 to ICANN on 30 April 1999, which – after an equally brief commenting period and a number of amendments99 – on 26 August 1999 adopted the UDRP.100 On 1 January 2000, a one-member panel decided the first domain name dispute under the UDRP, which has since governed well over 40,000 cases.101 The UDRP is a mandatory policy between a registrar and its customer and is included in registration agreements for all ICANN-accredited registrars. It thus applies for second-level domain name registrations in all generic Top Level Domains (gTLDs) (e.g., .com, .net and .org) and those country code top level domains (ccTLDs) that have elected to adopt the Policy.102 Under its mandatory administrative proceedings, ‘complainants’ (trademark owners) request cancellation, transfer or other changes to infringing domain name registrations in cases where (i) the domain name is identical or confusingly similar to a trademark or service mark in which the complainant has rights; (ii) if the domain name owner does not have rights or legitimate interests in respect of the domain name; and (iii) if the domain name has been registered and is being used in bad faith.103 Complaints are heard and decided by independent administrative panels (not the registrar)104 upon whose recommendation the registrars will cancel or transfer the domain name – thus providing a single system for adjudication and enforcement. This brief sketch of the UDRP’s genesis provides a glimpse of the complex interaction of interests that led to the creation of this unique regulatory 96

97

98

99 100

101 102 103 104

‘4) Nothing in the domain name registration agreement or in the operation of the new corporation should limit the rights that can be asserted by a domain name registrant or trademark owner under national laws.’ US White Paper (1998) Recommendations. The World Intellectual Property Organization (WIPO) publishes its Interim Report on the Internet Domain Name Process, 22 December 1998, Press Releases PR/1998/149. WIPO, Final Report of the WIPO Internet Domain Name Process, 30 April 1999, available at www.wipo.int/amc/en/processes/process1/report/finalreport.html. Komaitis (2010), p. 82. ICANN, Uniform Domain Name Dispute Resolution Policy (UDRP), adopted 26 August 1999, available at www.icann.org/resources/pages/schedule-2012-02-25-en. See www.wipo.int/wipo_magazine/en/2017/05/article_0008.html. Isenberg (2017). See www.icann.org/resources/pages/policy-2012-02-25-en. See paras 4(e) and (h), UDRP (n. 100).



 .     -                   

instrument.105 As the United States’ decision to privatize the DNS opened the registration process up to competition and multiple registrars, it became important to ensure the uniform ‘global’ application of dispute resolution procedures across different gTLDs.106 Early proposals to ensure this by requiring applicants to submit to personal jurisdiction at the place of the root ‘A’ server met with scepticism, as this was seen ‘as an inappropriate attempt to establish US trademark law as the law of the Internet’.107 Hence the realization in the US White Paper that accepting jurisdiction of ‘an alternative dispute resolution body is likely to be at least somewhat less controversial’.108 Finally, it was the common view among the engineering community that registrars’ involvement in resolving these disputes should be minimal. The UDRP clearly represents an international regulatory response to a digital IP challenge that differs entirely from the traditional treatymaking model. The unique constellation of a single global infrastructure – the DNS, that identifies servers and websites throughout the Internet – which is administered by a private corporation – ICANN, that is itself governed by a consensus-based multi-stakeholder model – created its own set of necessities and opportunities. The drive for a ‘supra-national’ uniform solution to domain name disputes was not motivated by aspirations of creating a ‘unified and universal body of rules’ for IP, but by the need – and the technical possibility – for uniform pragmatic solutions to a narrowly focused practical problem that occurred in the same fashion throughout the world, wherever a gTLD was registered. The UDRP is not an international treaty agreed by sovereign nations, nor even a national legislative act, but rather a set of standard terms that are included in the contractual relationship between applicant and registrar at the time of a gTLD registration. The single global infrastructure, and the individual contract with each domain name applicant during the registration process, permitted this unique contractual solution that enables conflict resolution and implementation in a single instance. This constellation

105

106

107

108

See the critical accounts of the processes involved in the privatization of the Internet and the creation of ICANN and the UDRP in Komaitis (2010). Chapter 5; Petillion and Janssen (2017), chapter 2; and Helfer and Dinwoodie (2001). The US Green Paper (1998) had still envisaged allowing different registrars to develop their own dispute resolution mechanisms. US White Paper (1998). The original root ‘A’ server, and the subsequent additional root servers were located in the United States. See ‘History of the Root Server System Report from the ICANN Root Server System Advisory Committee (RSSAC)’, 4 November 2016, available at www.icann.org/en/system/files/files/rssac-023-04nov16-en.pdf. US White Paper (1998).

       



is, however, also responsible for the narrow scope of application of the UDRP that focuses on providing a ‘quick and cheap’ solution to the most obvious cases.109 WIPO’s involvement provided a deliberative process with extensive involvement of WIPO member state governments, intergovernmental organizations, professional associations, corporations and individuals, as well as a panel of experts.110 This process – not least due to the time pressure under which it was conducted – cannot be compared to the iterative process of committee meetings that led to earlier international treaties,111 and was only the precursor to the subsequent internal ICANN stakeholder consultations and amendments of the Policy before its adoption in August 1999.112 While the UDRP has drawn repeated criticism – in terms of scope of application,113 procedural justice114 and judicial oversight115 – it is undisputed that this unique supra-national instrument has been a successful tool to resolve thousands of domain name disputes over the course of the last two decades.116 Although recourse to national courts remains a theoretical possibility under the Policy, the efficiency of the dispute resolution providers and the undeniable pecuniary advantages of arbitration means that few UDRP awards are challenged in courts.117

Regional Trade Agreements: Regulatory Responses to Digital Challenges Bilateral and regional treaties (collectively regional trade agreements or RTAs) that regulate aspects of IP law are not a new phenomenon.118 However, the recent dramatic increase in the number of those that address IP, and the degree of depth and detail of the IP provisions they 109 110 111

112 113 114 115 116 117

118

Dinwoodie (2000), p. 511. See WIPO Final Report (n. 98), paras 26–31 for the ‘Mechanics of the WIPO Process’. For doubts on WIPO’s ‘capacity to engage in a balanced consultative process’ under these circumstances, see Helfer and Dinwoodie (n. 105), p. 177. For a critical account of the ICANN process see above at n. 105. See Helfer and Dinwoodie (2001). See Helfer and Dinwoodie (2001), Geist (2001), Geist (2002). Petillion and Janssen (2017). de Werra (2016). See Petillion and Janssen (2017) and Helfer and Dinwoodie (2001). For a recent reversal of a UDRP award by the Paris Court of Appeal, see Pôle 5, chapter 1, Monsieur X./Team Reager AB et Stone Age Limited of 8 November 2016. For an account of bilateral IP treaties in force at the time of the conclusion of the Paris Convention, see Ricketson (2015), p. 2.02.



 .     -                   

80 19

70

60 22 50

24

9

40 30

6 3 7

2 3 5

23

22

15

19

16

20 10

19

24 7 6 6

16 12

11

7

0 Before 1995

1995–99

2000–04

RTAs without IP

Figure 4.1

2005–09

Negligible

2010–14

Moderate

2015–19 High

RTAs with increased IP content over time

Source: Valdés and McCann (2014) and WTO RTA Database.

contain, has made them a primary source of international regulatory responses to IP issues in recent decades (see Figure 4.1).119 Since 2000, IP-intensive RTAs have grown in a distinct hub-andspoke architecture – originally around the North American Free Trade Agreement (NAFTA), the European Free Trade Association (EFTA) and the European Union – and have encouraged the convergence of domestic IP regimes among the respective RTA signatories linked to a common hub.120 Higher standards agreed in such RTAs have fed back into the international arena, as countries tend to ‘re-export’ such commitments in subsequent RTAs in order to achieve deeper integration and to lock in domestic reforms.121 Implementing these commitments into domestic law usually leads to a non-discriminatory application of these new IP standards to all WTO member nationals, as the most-favoured nation (MFN) principle enshrined in the TRIPS Agreement does not permit general exceptions122 119

120 121 122

For a systematic quantitative analysis of IP provisions in RTAs, see Valdés and Tavengwa, (2012) and Valdés and McCann (2012). Valdés and McCann (2014) p. 36, and charts 6 and 7. Valdés and McCann (2014), p. 39. Article 4 of the TRIPS Agreement provides that ‘[w]ith regard to the protection of IP, any advantage, favour, privilege or immunity granted by a Member to the nationals of any other country shall be accorded immediately and unconditionally to the nationals of all other Members’. Exemptions exist only for general judicial assistance arrangements

       



for RTAs as are foreseen in GATT or GATS,123 where deeper integration in goods and services trade can be targeted exclusively to RTA partners. But even for provisions that fall outside the scope of the TRIPS MFN obligation,124 crafting national IP provisions that exclude third country nationals may be exceedingly costly and complex, considering that national treatment obligations are also contained in the Berne and Paris Conventions.125 Therefore, while RTAs themselves only bind their respective parties, the multilateral context and the practicalities of implementation favour the creation of a unilateral ‘international standard’ that RTA signatories apply to most, if not all, trading partners. The proliferation of RTAs with elaborate IP chapters leads to noteworthy results. Unlike the ‘spaghetti bowl’126 of preferential trading relationships created by the more permissive GATT and GATS rules, the ‘lasagne effect’127 of strict TRIPS MFN means that third countries outside the RTA can benefit from the higher IP standards that RTA signatories have negotiated among themselves, without having to provide the same level of protection. Third country nationals can benefit from higher IP standards agreed in an RTA – such as the availability of software patents or patent term extensions – by virtue of the TRIPS MFN and national treatment obligations without having to provide such standards

123

124 125 126

127

(Article 4(a)), protection dependent on reciprocity under the Berne or Rome Conventions (Article 4(b)), in respect of the rights of performers, producers of phonograms and broadcasting organizations not provided under this Agreement (Article 4(c), and deriving from notified international IP agreements that entered into force prior to the WTO Agreement (Article 4(d)). Contrary to GATT and GATS, the TRIPS Agreement does not contain a general MFN exemption for free trade agreements. See GATT Article XXIV and GATS Article V for exemptions from the respective MFN obligation for customs unions or free trade areas (or a corresponding interim agreement), or agreements liberalizing trade in services between or among parties, under certain conditions. See the limited exemptions from TRIPS MFN in Article 4(a)–(d), n. 122. See Article 5(1) Berne Convention (1971) and Article 2(1) Paris Convention (1967). The term was initially used by Bhagwati (1995) to denote the plethora of individual RTA trading relationships between WTO members that are permitted by the MFN exceptions under GATT and GATS, and thus create a complicated web of relationships much like a spaghetti bowl. See Taubman (2008) and Taubman (2019). The absence of general exceptions from the TRIPS MFN obligation means that additional substantive obligations agreed in an RTA lead to a general ‘ratcheting up’ of that country’s IP obligations vis-à-vis nationals of all other WTO members. Thus, in contrast to the situation of the GATT/GATS spaghetti bowl described at n. 126, IP commitments in RTAs add general layers to the TRIPS obligations of that member, much like – to extend Bhagwati’s food metaphor – in a lasagne dish.



 .     -                   

to nationals of the RTA signatories. This means, inversely, that the value of accepting an IP commitment in RTA negotiations – for instance in exchange for market access in a certain agriculture sector – should take into account its non-discriminatory scope of application. While – in the sphere of trade in goods – the benefits of a lower tariff for beef or rice can be limited to an RTA trading partner in accordance with the exceptions to the MFN principles under the GATT, the situation is different under TRIPS. The more limited scope of exceptions to MFN in TRIPS means that bilateral IP commitments such as specific digital enforcement standards will have to be made available to nationals of all members; and their scope and impact is thus more difficult to ascertain. Against that background, the regulatory impact of provisions in RTAs goes well beyond the jurisdictions of signatories and has contributed significantly to an internationalization of certain approaches to emerging IP issues – including in the digital sphere. As the frequency and complexity of IP chapters in RTAs grow at dramatic speed, detailed qualitative analysis of IP commitments remains a challenge. Quantitative analysis shows a broad distribution of IP topics addressed in RTAs (see Figure 4.2). RTAs frequently address digital IP issues concerning copyright law, trademark law, IP enforcement, and technology-specific, trade-related matters such as the ISP liability. RTA commitments on copyright include clarifying the application of existing remedies to online enforcement or undertaking to target specific copyright violations in the online environment.128 Many IP chapters refer to the WIPO Internet

128

Example: China–Korea, Republic of: Section J, Article 15.28 Each Party shall take effective measures to curtail repetitive infringement of copyright and related rights on the Internet or other digital network. . . . Section K, Article 15.30: Under the established structure of this Agreement, each Party shall, upon request of the other Party, and in addition to the already existing forms of cooperation: . . . (ii) exchanges and cooperation on online copyright enforcement Side letter – Example: Korea, Republic of–US: Confirmation letter (Online Piracy Prevention): The Parties agree on the objective of shutting down Internet sites that permit the unauthorized reproduction, distribution, or transmission of copyright works, of regularly assessing and actively seeking to reduce the impact of new technological means for committing online copyright piracy, and of providing generally for more effective enforcement of IPRs on the Internet. The Republic of Korea agrees that internet piracy of works and other subject matter protected by copyright . . . The Republic of Korea agrees to issue as soon as possible, but no later than six months after the date the Agreement enters into force, a policy directive establishing clear jurisdiction for a division or joint investigation team to engage in effective enforcement against online piracy. . .

       



Geographical indications

Trademarks Copyright and related rights Patents Plant varieties Industrial designs Trade secrets Traditional knowledge/Genetic resources Encrypted signals Domain names 0%

Figure 4.2 Specific IP provisions in RTAs

10% 129

20%

30%

40%

50%

60%

with IP

Source: Valdés and McCann (2014) and WTO RTA Database.

Treaties,130 variously requiring compliance with, affirming existing obligations pursuant to and requiring accession to these treaties.131 Some of the most detailed copyright provisions concern technical protection measures (TPMs), reflecting the considerable variety in national implementation of the general obligation to provide for TPMs under the WIPO Internet Treaties. Notably RTAs negotiated with the participation of the United States contain detailed provisions in this regard that may define the subject matter of TPM protection,132 require protection 129 130 131

132

RTAs that have been notified to the WTO and are in force 2019. See Valdés and McCann (2014), p. 15. Example: EFTA–Peru: Chapter 6, Article 6.4.3 The Parties to this Agreement which are not a party to one or more of the agreements listed below shall ratify or accede to the following multilateral agreements within one year from the date of entry into force of this Agreement: (a) WIPO Performances and Phonograms Treaty of 20 December 1996 (WPPT); and (b) WIPO Copyright Treaty of 20 December 1996 (WCT). Example: CETA: Article 20.7.1 The Parties shall comply with the following international agreements: (a) Articles 2 through 20 of the Berne Convention for the Protection of Literary and Artistic Works, done at Paris on 24 July 1971; (b) Articles 1 through 14 of the WIPO Copyright Treaty, done at Geneva on 20 December 1996; (c) Articles 1 through 23 of the WIPO Performances and Phonograms Treaty, done at Geneva on 20 December 1996; and [. . .]. Example: Korea, Republic of–US: Article 18.4.7 (f ). Effective technological measure means any technology, device, or component that, in the normal course of its operation, controls access to a protected work, performance, phonogram, or other protected subject matter, or protects any copyright or any rights related to copyright.



 .     -                   

of TPMs independently of any underlying copyright infringement133 and sometimes specify the legal remedy that must be available against the circumvention (i.e. ‘hacking’) of such TPMs.134 In the area of trademark law, RTA commitments frequently refer to the 2001 WIPO Joint Recommendation on marks and signs on the Internet (JR 2001), either affirming its importance, or providing hortatory language on applying it.135 As mentioned above, this mechanism has helped spread and multiply pledges to apply the non-binding JR 2001 among RTA signatories and thus helped – to some extent – turn soft into hard law among RTA signatories. Many RTAs also contain provisions on domain names and trademark use on the Internet that address the bad-faith registration or use of a domain name that is confusingly similar to a trademark or, in the case of EU – CARIFORUM States the Economic Partnership Agreement (EPA), a trademark or a geographical indications (GI).136 Some of these provisions make explicit reference to ICANN’s UDRP as a model for how such conflicts should be resolved. Finally, on enforcement, RTA provisions range from general commitments on cooperation and best efforts to address Internet piracy, to

133

134

135

136

Example: US–Singapore: Article 16.4.7(d). Each Party shall provide that a violation of the law implementing this paragraph is independent of any infringement that might occur under the Party’s law on copyright and related rights. Example: CPTPP: Chapter 20, Article 20.66: Technological Protection Measures 1. In order to provide adequate legal protection and effective legal remedies against the circumvention of effective technological measures that authors, performers, and producers of phonograms use in connection with the exercise of their rights and that restrict unauthorized acts in respect of their works, performances, and phonograms, each Party shall provide that a person who: (a) knowingly, or having reasonable grounds to know, circumvents [. . .]; or (b) manufactures, imports, distributes, offers for sale or rental to the public, or otherwise provides devices, products, or components, or offers to the public or provides services, that: [. . .] is liable and subject to the remedies provided for in Article 20.81.18 (Civil and Administrative Procedures and Remedies). See US–Singapore: Article 16.2(b) (shall give effect to JR 1999) and EU –CARIFORUM, Article 145 (shall endeavour to apply JR 2001). Example: CPTPP: Chapter 18, Article 18.28.1 In connection with each Party’s system for the management of its country-code top-level domain (ccTLD) domain names, the following shall be available: (a) an appropriate procedure for the settlement of disputes, based on, or modelled along the same lines as, the principles established in the Uniform Domain-Name Dispute-Resolution Policy, as approved by the Internet Corporation for Assigned Names and Numbers (ICANN). 2. In connection with each Party’s system for the management of ccTLD domain names, appropriate remedies shall be available at least in cases in which a person registers or holds, with a bad faith intent to profit, a domain name that is identical or confusingly similar to a trademark. See also EU–CARIFORUM Article 145.

       



providing specific remedies or adopting specific interpretations of elements for criminal liability.137 Among the most elaborate provisions in this area are commitments relating to the liability of ISPs for IP infringements committed by their users, or via their platforms, which are discussed in detail below.138 The impact of bilateral treaties on the international normative landscape of IP regulation has produced a considerable body of literature by academia and international organizations over recent decades139 which provide a rich and detailed discussion of the strategic and institutional particularities involved in their conclusion. The present chapter concentrates on several distinctive characteristics of RTAs as international regulatory responses to emerging digital challenges. The inclusion of IP chapters in RTAs does not directly advance the goal of a unified body of IP rules for universal application, as did the early international IP conventions. Nevertheless, most of these IP chapters are expressly situated in the context of the multilateral IP treaties by reaffirming their importance, exhorting compliance with them, or specifically recognizing their precedence over the RTA rules.140 The focus of RTA IP chapters – when they venture beyond general affirmations – is on regulating IP areas or issues outside the scope of the multilateral rules, or where those rules provide discretion for national implementation. In some cases, this will mean inciting partners to adopt international standards they had hitherto not joined on their own account, thus reinforcing linkages to the international IP treaties and increasing their membership. In areas where no examples of international codification exist, the dynamic of RTA IP negotiations will typically be to assess whether to resist or accept demands for adoption of particular regulatory IP approaches that already exist in one of the RTA partners – leading to what is often referred to as ‘transplantation’.141 137

138 139 140

141

See digital enforcement measures contained in Australia–Chile; Dominican Republic– Central America–United States Free Trade Agreement (CAFTA-DR); EC; EU– Colombia; EU–Republic of Korea; EU–Peru; Japan–Peru; Japan–Switzerland; Republic of Korea–United States; Mexico–Nicaragua; North American Free Trade Agreement (NAFTA); United States–Australia; United States–Bahrain; United States–Chile; United States–Colombia; United States–Morocco; United States–Panama; United States–Oman; United States–Peru; United States–Singapore. Section 3.1.2. See Seuba (2015) and note 8. See examples for TRIPS and WIPO treaties’ affirmation in RTAs in Valdés and McCann (2014), paras 45–54. Valdés and McCann (2014), para. 150. Seuba (2015), p. 63.



 .     -                   

The distribution of interests in these negotiations is distinct from international rule-making not only because of the limited number of participants – where differences in negotiating capacity take on greater significance – but also because RTAs are typically negotiated across a number of trade-related areas, inviting linkages and trade-offs across different subject areas that have been ostensibly absent, or at least more difficult to bring to bear, in traditional international negotiations that have been focused on IP. These elements also impact on the degree of expert participation in these negotiations, a factor which had arguably enabled the traditional international treatymaking process to develop or adapt original solutions to specific IP issues, such as in the WCT and the WPPT. The time pressure associated with RTA negotiations and the constellation of interests described above indicates that the predominant perspective in RTA IP negotiations is the assimilation of systems by adoption, not necessarily the development of new substantive solutions to IP challenges. It seems that it is only in negotiations between partners with strongly divergent interests in a particular area that RTA negotiators attempt to craft creative solutions espousing common principles.142

Conclusion The mechanisms of international regulatory responses to emerging IP issues have evolved significantly in the digital era. For most of the twentieth century the institutionalized deliberative process of the traditional treatymaking model provided an ecosystem in which systemic considerations and comparative assessment enabled a broader and potentially more comprehensive assessment of regulatory solutions that often led to a common, and sometimes original, multilateral outcome. The arrival of digitization and online distribution represents a significant acceleration and proliferation of technological change. The corresponding speed and urgency of regulatory reaction could be argued to have tilted the balance – also in the international process described above – towards demands for individual short-term responses which are difficult to resist, thus reflecting the realisation that ‘it is easier to reach a shared understanding on specific issues, where there is a demonstrable and manageable need for international action, than to achieve a 142

See the provisions on geographical indications in CETA, US–Republic of Korea and the United States–Mexico–Canada Agreement (USMCA).

       



shared understanding across the whole range of IP, which now underlies most economic and cultural activities’.143 Thus, coinciding with the advent of the digital era, this traditional treatymaking scenario has given way to more varied landscape where a variety of tools and legal instruments embody regulatory responses to digital challenges. The preceding section sought to identify the factors that shape different regulatory responses to new issues.

Ready for a Common Regulatory Response? – Selected Emerging Digital Issues In the global, potentially seamless markets for digital products that have been created by digital communication technologies, common – or at least interoperable – regulatory solutions to IP issues are essential to ensure the viability of digital business models through which much of the continuously increasing share of online trade is conducted. This part briefly surveys several salient digital IP issues with regard to an available international regulatory response.

ISP Liability The degree of liability of ISPs for IP infringements by their users directly impacts the viability of digital platforms in a global market. A common international approach to this question would therefore be particularly valuable. Indeed, the pervasive view that the important role in facilitating access justifies a certain privilege for ISPs, requiring their action and collaboration only when they are notified of infringing content by rights holders, was developed through years of jurisprudence and – once the ‘right of making available’ had been established by the WCT and WPPT144 – led to national legislative responses centred on similar, yet not identical, principles.

Regulatory Approaches in National Legislation Two early legislative implementations of the WCT and the WPPT, the US Digital Millennium Copyright Act (DMCA) (1998) and the European 143

144

See ‘Acceptance Speech by Francis Gurry on his Reappointment’, WIPO Assemblies 2014, 8 May 2014, available at www.wipo.int/about-wipo/en/dgo/speeches/a_53_dg_ speech.html. See ‘The WIPO Internet Treaties: An Issue-Driven International Solution’ above.



 .     -                   

E-Commerce Directive (2000), enshrined systems of ISP liability limitations into national law. Their evolution and case law – from the perspective of music streaming – are the subject of Chapter 19 of this volume.145 This chapter identifies aspects that illustrate similarities and differences with regulatory responses taken in other jurisdictions. United States and European Union The US DMCA146 establishes limitations on the copyright liability of online service providers147 that maintain a policy providing for termination of network services for repeat infringers in appropriate circumstances148 and do not interfere with standard technical measures used by copyright owners to identify and protect copyrighted works.149 For such service providers, ‘safe harbours’ are available for transitory communications,150 system caching,151 storage of information on systems or networks at the direction of users (i.e. hosting),152 and information location tools (i.e. search engines).153 A notice and take-down procedure154 prescribes the conditions under which the ISP can be considered to have actual knowledge of infringing material on its services, and monetary liability is waived, if after receiving the notification, the ISP quickly removes or blocks access to the material identified in the notification,155 while the subscriber has the possibility of

145 146 147

148 149 150

151

152 153 154 155

LaFrance (2020), Chapter 19 in this volume. Public Law 105–304, 28 October 1998. Service providers are defined in section 512(k)(1)(A) as ‘an entity offering the transmission, routing, or providing of connections for digital online communications, between or among points specified by a user, of material of the user’s choosing, without modification to the content of the material as sent or received’. 17 USC § 512(i)(1)(A). 17 USC § 512(i)(1)(B). 17 USC § 512 (a). Transitory Communications are activities where the service provider acts as a ‘mere conduit’ of digital information from one point to another including transmissions, routing and connection services. 17 USC § 512 (b). System caching is the activity carried out by a service provider, when it retains, for limited time, copies of material that has been uploaded by a third party to be transmitted to an Internet user, in order to simplify interactions online and reduce the service provider bandwidth requirements. 17 USC § 512 (c). 17 USC § 512 (d). 17 USC § 512(c)(3) for hosting services and 512(d) for information location tools. As an additional safeguard for the ISP for this diligent action, section 512(g)(1) establishes that there will be no liability to any third party if the ISP took down allegedly infringing material in ‘good faith’.

       



counter-notification156 to respond to any erroneous or fraudulent notifications. The act permits injunctions against ISPs that may include ordering the ISP to restrain from providing access to the infringing material or activity, thus restraining the service provider from providing access to a subscriber who is engaging in infringing activity and is identified in the order or any other injunctive relief necessary to prevent infringement.157 This DMCA approach to ISP liability, as developed and confirmed in subsequent case law, set a relatively high threshold for knowledge of the ISP necessary to trigger its obligation to act,158 and has expressly excluded an obligation to monitor or filter content on its own accord,159 that has recently come under criticism.160 In the European Union, broad exemptions from ISP liability – covering also IPRs beyond copyright, as well as unfair competition and criminal acts – were established in the E-commerce Directive.161 Under the Directive an information society service provider can enjoy exemption from liability for mere conduit,162 caching163 and hosting164 of third-party 156

157 158

159

160

161

162

163 164

512(g)(1). If a counter notification is filed in compliance with the requirements, and the copyright owner does not file an action seeking court order against the subscriber, the service provider has to re-grant access to the material within 10–14 business days after receiving the notification. Knowing misrepresentation that material is an infringement incurs liability for subsequent injury (section 512(f )). Section 512(j). See the discussion of ‘actual’ and ‘red flag’ knowledge in Viacom Int’l, Inc. v. YouTube, Inc., 676 F.3d 19 (2nd Cir. 2012); Capitol Records, LLC v. Vimeo, LLC, 972 F. Supp. 2d 500, 523 (SDNY 2013); Capitol Records, Inc. v. MP3tunes, LLC, 821 F. Supp. 2d 627, 633–634 (SDNY 2011) (‘MP3tunes I’). See also the references in LaFrance (n. 145). Section 512(m)(1). See also Viacom Int’l, Inc. v. YouTube, Inc., 676 F.3d 19 (2nd Cir. 2012) and Capitol Records, LLC v. Vimeo, LLC, 972 F. Supp. 2d 500, 523 (SDNY 2013). See Artists Urging Reforms of the DMCA Safe Harbor: ‘Our Culture Is At Stake’ (14 March 2017), available at https://blog.jipel.law.nyu.edu/2017/03/artists-urging-reformsof-the-dmca-safe-harbor-our-culture-is-at-stake/. Directive 2000/31/EC of the European Parliament and of the Council of 8 June 2000 on certain legal aspects of information society services, in particular electronic commerce, in the Internal Market (‘Directive on electronic commerce’). Article 12 E-commerce Directive provides that ‘where an information society service is provided that consists of the transmission in a communication network of information provided by a recipient of the service, or the provision of access to a communication network’, ISPs will not be liable if the ISP does not initiate the transmission; the ISP does not select the receiver of the transmission; and the ISP does not select or modify the information contained in the transmission. Article 13. Article 14.



 .     -                   

information. Exemption from liability for caching and hosting is conditional on the service provider acting expeditiously to remove or to disable access to the information, upon obtaining ‘actual knowledge or awareness’ of illegal activity or information.165 Under the E-commerce Directive, member states were expressly prohibited from imposing a general monitoring obligation on ISPs,166 although monitoring obligations under national law ‘in a specific case’ – arguably including content already identified as infringing (the so-called ‘stay down’ obligation) – were expressly contemplated.167 The 2001 Directive on Copyright in the Information Society (InfoSoc Directive)168 that implemented the WCT and WPPT did require that injunctions can be addressed to ISPs ‘whose services are used by a third party to infringe a copyright or related right’,169 but otherwise left the remaining question of what liability ISPs retained in the area of copyright infringements largely to member states.170 Characteristically, these early systems focused on establishing the conditions for liability exemptions for ISPs, rather than clarifying the boundaries of (primary or secondary) copyright liability of ISP for their activities. The specific ISP liability for injunctions to take down or disable access to infringing material – available under both the DMCA and the E-commerce directive171 – was a recognition of accountability for terminating, not liability for committing,172 third-party infringements and was not systemically embedded to the overall conceptual regime of

165 166

167

168

169 170 171

172

For caching, see Article 13(e). For hosting, see Article 14. Article 15.1 ‘Member States shall not impose a general obligation on providers, when providing the services covered by Articles 12, 13 and 14, to monitor the information which they transmit or store, nor a general obligation actively to seek facts or circumstances indicating illegal activity.’ Recital 47: (47) ‘Member States are prevented from imposing a monitoring obligation on service providers only with respect to obligations of a general nature; this does not concern monitoring obligations in a specific case and, in particular, does not affect orders by national authorities in accordance with national legislation.’ Directive 2001/29/EC of the European Parliament and of the Council of 22 May 2001 on the harmonization of certain aspects of copyright and related rights in the information society. Article 8.3 InfoSoc Copyright Directive. EU Parliament (2018). For the EU, see Article 8(3) Copyright Directive (for copyright) and Article 11, 3rd sentence Enforcement Directive (for other IPRs). For the US, see DMCA section 512(j). EU Parliament (2018), 20, ‘Precisely put, Article 8(3) Copyright Directive is not about a provider liability, but helping duties. Husovec called this “accountable, not liable”.’ See Husovec (2016).

       



liability for copyright infringement. It reflected the then current perception of intermediaries as passive participants in the transmission of information, without any technical capacity to monitor third-party content,173 who are nonetheless well placed to help in terminating the infringements of other – third-party – infringers.174 There has thus developed a rich body of subsequent jurisprudence and debate to calibrate the exact balance of obligations between right owners and ISPs, to ensure systemic consistency with the concepts of primary and secondary liability for copyright infringements and to take account of the dramatically increased economic impact and technological ability of ISPs today.175 In the US context, where monitoring duties of ISPs had been categorically excluded by legislation, the main area of engagement was the precise contours of the knowledge – ‘red flag’ or actual – required to trigger the ISPs’ obligations to take action and how primary liability could be established for ISPs falling outside the safe harbour.176 In the European Union, this also included the contemplation of certain ISP monitoring duties – which remained available under national systems – a thorough debate of weighing different fundamental rights and more recently, the ‘value gap’ discussion.177 Although these two early systems of ISP ‘safe harbours’ were conceptually very similar, their subsequent reception in jurisprudence and debate has clearly pulled in different directions. Despite growing criticism, US jurisprudence interpreting the DMCA has maintained a broad scope of the ‘safe harbour’ and a high threshold for triggering ISP liability for removal of infringing material. In stark contrast, in the European Union, the contemplation of certain monitoring duties178 and

173 174

175 176

177 178

Rotaru (2017). See InfoSoc Directive, Recital 59 in fine: ‘In the digital environment, in particular, the services of intermediaries may increasingly be used by third parties for infringing activities. In many cases such intermediaries are best placed to bring such infringing activities to an end.’ LaFrance (2020), Chapter 19 in this volume. See the discussion of ‘actual’ and ‘red flag’ knowledge in Viacom Int’l, Inc. v. YouTube, Inc., 676 F.3d 19 (2nd Cir. 2012); Capitol Records, LLC v. Vimeo, LLC, 972 F. Supp. 2d 500, 523 (SDNY 2013); Capitol Records, Inc. v. MP3tunes, LLC, 821 F. Supp. 2d 627, 633–634 (SDNY 2011) (‘MP3tunes I’). See also LaFrance (2020), Chapter 19 in this volume. See the succinct overview of the value gap argument in Senftleben (2020). See EU Parliament (2018) p. 17 on the lack of decisive CJEU jurisprudence on the scope of Article 15 E-Commerce Directive.



 .     -                   

the Court of Justice of the European Union’s (CJEU’s) development of sector-specific liability rules for the right of making available179 prepared the ground for a significant reversal of direction in the 2019 Directive on Copyright in the Digital Single Market (DSMD)180 which European member states will have to implement by 2021. Significantly, the DSMD establishes primary copyright liability for large commercial platforms carrying user-generated content181 for unauthorized acts of making available or communication to the public unless best efforts were made to obtain authorization, to ensure unavailability (i.e. monitors and filters) of specific identified works in accordance with high industry standards of professional diligence, and operates a ‘notice and stay down’ system – all this while maintaining availability of content which is – including by virtue of exceptions such as criticism or parody – non-infringing.182 This paradigm-shifting rebalancing of the burden of monitoring the Internet for copyright violations – from the right owners to the hitherto more privileged ISPs – is setting ISP liability on a new course in the European Union, and is already shifting the debate towards the merits of the expected large-scale application of automated algorithmic enforcement.183 Non-Signatories of the WIPO Internet Treaties The early systems of safe harbours for ISP liability set up by the US and the EU, including the subsequent jurisprudence and debate, have also influenced national regulatory approaches in other jurisdictions – including in nonsignatories of the WIPO Internet Treaties, which did thus not necessarily have the ‘making available’ right defined in their legislation. They also sought to balance the obligations of right owners and ISPs concerning online copyright infringement.

179

180

181

182 183

CJEU of 8 September 2016, C-160/15 – GS Media/Sanoma; CJEU of 26 April 2017, C-527/15 – Brein/Wullems (Filmspeler); CJEU of 14 June 2017, C-610/15 – Ziggo/Brein. Directive (EU) 2019/790 of the European Parliament and of the Council of 17 April 2019 on copyright and related rights in the Digital Single Market and Amending Directives 96/9/EC and 2001/29/EC, O.J. (L 130) 92. See Article 17(1) and Article 17(6) for the definition of ‘online content-sharing service provider’ and the scope of application. Articles 17(4) and 17(7). See Senftleben (2020) and Schwemer and Schovsbo (2020).

       



India India, a non-signatory of the Internet Treaties until 2018,184 introduced successive ISP safe harbours in a 2008 amendment to its Information Technology Act 2000 (IT Act), and in a 2012 amendment of its Copyright Act 1957. The 2008 amendment of the IT Act introduced a broad liability exemption for intermediaries whose functions were limited to providing access to a communication system as long as they observed ‘due diligence’ in discharging their functions, and expeditiously removed or disabled access to the offending material upon receiving actual knowledge or on being notified by a government agency.185 Uncertainty on the extent of ‘due diligence’, and on the broader question of whether this safe harbour would be available also regarding copyright violations led to jurisprudence contemplating broad monitoring duties for ISPs. In Super Cassettes Industries Limited v. MySpace Inc.,186 the Single Judge Bench of the Delhi High Court (HC) granted an interim injunction on the grounds that there was no safe harbour for intermediaries under the Copyright Act187 and implied that intermediaries should screen all user-generated content to check for infringement prior to making the content available online.188 Although the Division Bench upheld the eligibility of the ISP for safe harbour under the IT Act,189 uncertainty remains around how to ensure harmonious interpretation of the different statutes in different scenarios, and what degree of knowledge is necessary to trigger ISP action. A copyright-specific safe harbour was added in the 2012 amendment of the Copyright Act, which expressly pursued the aim of implementing the Internet Treaties.190 It introduced an exemption from secondary liability for copyright infringement for ‘transient or incidental storage’ of material unless the ISP is ‘aware or has reasonable grounds’ for believing that such storage is of an infringing copy.191 A notice and take-down procedure requires that access must be withheld upon

184

185 186

187 188 189 190 191

India acceded to the Internet Treaties on 25 September 2018. The treaties entered into force for India on 25 December 2018. See Articles 2(w) and 79(1)–(3) Information Technology Act (2000) as amended. In Super Cassettes Industries Limited v. MySpace Inc. & Another, IA Nos. 15781/2008 &3085/2009 in CS (OS) No. 2682/2008. Agarwal and Agarwal (2017). Kumar (2014). Agarwal and Agarwal (2017), 187. See statement by the Indian Copyright office, available at http://copyright.gov.in/. 52.1(c) Copyright Act 1957, as amended.



 .     -                   

receiving a written complaint from the right owner for 21 days, awaiting a court order.192 The coexistence of these two Indian safe harbour regimes, in s. 79 of the IT Act and in s. 52.1(c) the Copyright Act, for which jurisprudence has developed different knowledge standards193 and whose respective scope of application have not been clearly delineated194 means that concrete litigation outcomes continue to rely on courts applying and interpreting principles of copyright liability and statute concordance, rather than on a predictable tool to preserve ISP business models. Brazil Not being a signatory to the Internet Treaties, judges and prosecutors in Brazil had to identify and interpret national law in order to encompass the concept of ‘making available’ by platforms and ISPs. In the absence of specific rules before 2014, the application of general principles of civil law akin to strict liability had led to a number of private agreements between copyright holders and ISPs that established voluntary notice and take-down procedures that reflected industry demands.195 Against this background, the vigorous discussion around a controversial far-reaching 2008 proposal regarding internet activities more generally,196 followed by an elaborate public consultation during 2008–2009 and a protracted legislative process, led to the adoption of the 2014 Law No. 12.965, also known as the ‘Marco Civil da Internet’197 which sets out the regulatory framework for internet activities, and presents a complex compromise with regard to ISP liability. ISP activities of ‘mere conduit’ are exempted from liability198 without the conditions required in other jurisdictions. Internet application providers (i.e. content hosts) are exempt if they remove content upon receipt of a specific judicial order.199

192 193 194

195 196

197

198 199

52.1(c) Copyright Act 1957, as amended. Sethia (2017), Saha and Saha (2018). Agarwal and Agarwal (2017). See also Vodafone India Ltd v. RK Productions, 2013 (54) PTC 149 (Mad), in which the Madras High Court upheld the application of s. 79 IT Act to copyright infringement. Zingales (2015). The so-called ‘Azeredo Bill’ proposed three-year mandatory data retention and other duties for ISPs, as well as the criminalization of access to data ‘without authorization of the legitimate owner’. See Zingales (n. 195). Law No. 12.965 of April 23, 2014, available in English at www.publicknowledge.org/ assets/uploads/documents/APPROVED-MARCO-CIVIL-MAY-2014.pdf. Article 18. Article 19.

       



Only material that constitutes a breach of privacy200 must be removed at the simple – extra-judicial – requests of an interested party.201 While the retention of a judicial notice as a requirement for ISP action was seen as a success of the civil rights movement,202 this was achieved at the cost of a carve-out of liability for copyright and related rights violations, which is excluded from the safe harbour provisions described above pending specific legislation.203 The National Congress’s discussions of an amendment of the Copyright Bill currently contemplate the instruments of ‘notice and notice’ – an obligation on the ISP to notify the alleged infringer upon receiving a notification by the right holder, and a proposal to provide right holders with a remuneration right for each improper use.204 In the absence of a specific law as required by the statute, courts continue to rely on general principles of secondary copyright liability, under which the current practice remains of the removal of infringing material upon extra-judicial notification by the right holder.205 The above examples illustrate areas of commonality, but also of considerable variety that continues to exist in national ISP safe harbour regimes. Common ground seems to exist for establishing the susceptibility of ISPs to injunctions for removing access to infringing material, although there is considerable diversity as to whether such liability is rooted in secondary liability for copyright infringement or whether it is a general ‘helping duty’ due to an ISP’s practical ability to terminate infringement by a third party. A certain commonality also exists with respect to the types of ISP activities that may benefit from a safe harbour, with caching and hosting being identified in most regimes, while location services are already more infrequently covered. Beyond these – rather basic – concepts, considerable variety exists with respect to the degree of ISP knowledge required to trigger cooperation, the nature of the ISP’s cooperation that might be triggered, and, in particular, the question of how ISP activity outside the safe harbour would incur liability for copyright infringement. 200 201 202 203

204 205

Essentially private acts of nudity or sexual activity (also known as ‘revenge-porn’). Article 21. Mulholland (2018). See Article 19.2 ‘Application of the provision in this article to infractions of copyright or related rights depends on a specific legal provision, which must respect freedom of expression and other guarantees set forth in article 5 of the Federal Constitution.’ Mulholland (2018) at 1.9. Ibid.



 .     -                   

The latter question is further complicated by the rapid economic and technological development of Internet enterprises that provide access and host content, representing business models that have long surpassed the original perception of passive access providers with little technological means to control or monitor the content to which they provide access.

Provisions in Regional Trade Agreements Considering the interest of Internet companies in minimizing the cost of compliance with diverse liability regimes in different markets – and reflecting the role of RTAs described earlier in this paper – it is no surprise that provisions on ISP liability occur with increasing frequency and depth in recent RTAs (Table 4.1). In line with the purpose of this chapter, the following examples are by no means exhaustive but attempt to briefly illustrate the assimilation of regulatory responses in the field of ISP liability under current RTAs. As is to be expected from the hub-and-spoke architecture described above, the RTAs with the US and the EU largely reflect their domestic systems and approaches to the issue. While early agreements with these partners may contain individual provisions regarding ISPs,206 more elaborate RTAs such as CETA, EU–Republic of Korea, US–Singapore, US–Republic of Korea, TPP and USMCA contain elaborate specific obligations that essentially mirror the entire domestic EU and US systems described in the previous section. Beyond the EU and US agreements, general provisions regarding ISP cooperation in RTAs take the form of permissive contemplation of limiting ISP liability207 or individual provisions requiring that user identity can be obtained from the ISP in the context of Internet copyright enforcement,208 or that ISP liability shall be limited to encourage cooperation with right holders.209

206

207

208 209

See, for example, EU–CARIFORUM States EPA: Section 2, Article 158 (permit injunctions against ISPs). ‘China–Australia Article 11.20: Service Provider Liability Each Party may take appropriate measures to limit the liability of, or remedies available against, internet service providers for copyright infringement by the users of their online services or facilities, where the internet service providers take action to prevent access to the materials infringing copyright in accordance with the laws and regulations of the Party.’ China–Republic of Korea, Article 15.29 and Japan–Switzerland, Article 126.2. Japan–Switzerland, Article 126.1.

Table 4.1 Selected ISP provisions in selected EU and US RTAs

Liability of ISPs



EU–Republic of Korea

CETA

US–Republic of Korea

USMCA

Article 10.62

Article 20.11(1)

Article 18.10(30)

The Parties recognise that the services of intermediaries may be used by third parties for infringing activities. To ensure the free movement of information services and at the same time enforce intellectual property rights in the digital environment, each Party shall provide for the measures set out in Articles 10.63 through 10.66 for intermediary service providers where they are in no way involved with the information transmitted.

Subject to the other paras of this Article, each Party shall provide limitations or exceptions in its law regarding the liability of service providers, when acting as intermediaries, for infringements of copyright or related rights that take place on or through communication networks, in relation to the provision or use of their services.

1. For the purpose of providing enforcement procedures . . . each Party shall provide . . . : (a) legal incentives for service providers to cooperate with copyright [fn 35: and related rights] owners in deterring the unauthorized storage and transmission . . .; and (b) limitations in its law regarding the scope of remedies available against service providers for copyright infringements that they do not control, initiate, or direct, . . . .[fn 36: without prejudice to general defences for copyright infringement].

Article 20.88: Legal Remedies and Safe Harbors 1. The Parties recognize the importance of facilitating the continued development of legitimate online services operating as intermediaries and, in a manner consistent with Article 41 of the TRIPS Agreement, providing enforcement procedures that permit effective and expeditious action . . . This framework of legal remedies and safe harbors shall include: (a) legal incentives for Internet Service Providers to cooperate with copyright owners

Table 4.1 (cont.) EU–Republic of Korea

CETA

USMCA

Article 18.10(30)

to deter the unauthorized storage and transmission of copyrighted materials . . .; and (b) limitations in its law that have the effect of precluding monetary relief against ISPs for copyright infringements that they do not control, initiate or direct . . . Article 20.88

1.(b) . . . (i) These limitations shall preclude monetary relief and provide reasonable restrictions on courtordered relief . . . for the following functions, and shall be confined to those functions:

2. The limitations described in para. 1(b) shall include limitations in respect of the following functions: (a) transmitting, routing, or providing connections for



US–Republic of Korea

ISP functions covered

Article 10.63 ‘mere conduit’

Article 20.11(2) The limitations . . .

1. Where . . . service . . . consists of the transmission in a communication network . . . Article 10.64 ‘caching’ 1. Where an information society service . . .

(a) shall cover at least the following functions: (i) hosting of the information at the request of a user of the hosting services; (ii) caching carried out through an automated



consists of the transmission in a communication network of information provided by a recipient of the service, . . . Article 10.65 ‘hosting’ 1. Where an information society service . . . consists of the storage of information provided by a recipient of the service, the Parties shall ensure that the service provider is not liable for the information stored . . .

process, when the service provider: . . . (iii) mere conduit, which consists of the provision of the means to transmit information provided by a user, or the means of access to a communication network; and (b) may also cover other functions, including providing an information location tool, . . .

(A) transmitting, routing, or providing connections for material without modification of its content, or the intermediate and transient storage of such material in the course thereof; (B) caching carried out through an automatic process; (C) storage at the direction of a user of material residing on a system or network controlled or operated by or for the service provider; and (D) referring or linking users to an online location by using information location tools, including hyperlinks and directories.

material without modification of its content or the intermediate and transient storage of that material done automatically in the course of such a technical process; (b) caching carried out through an automated process; (c) storage, at the direction of a user, of material residing on a system or network controlled or operated by or for the Internet Service Provider; and (d) referring or linking users to an online location by using information location tools, including hyperlinks and directories.

Table 4.1 (cont.)

Conditions to qualify for limitation of liabilty



EU–Republic of Korea

CETA

US–Republic of Korea

USMCA

Article 10.66 (2)

Article 20.11(4)

Article 18.10(30)

Article 20.88

The Parties may establish obligations for information society service providers to promptly inform the competent authorities of alleged illegal activities undertaken or information provided by recipients of their service, or to communicate to the competent authorities, at their request, information enabling the identification of recipients of their service with whom they have storage agreements.

Each Party may prescribe in its domestic law, conditions for service providers to qualify for the limitations or exceptions in this Article. Without prejudice to the above, each Party may establish appropriate procedures for effective notifications of claimed infringement, and effective counternotifications by those whose material is removed or disabled through mistake or misidentification.

1.(b)(v) With respect to functions referred to in clauses (i)(C) and (D), the limitations shall be conditioned on the service provider: (A) not receiving a financial benefit directly attributable to the infringing activity, in circumstances where it has the right and ability to control such activity; (B) expeditiously removing or disabling access to the material residing on its system or network on obtaining actual knowledge of the infringement or becoming aware of facts or circumstances from which the infringement

3. To facilitate effective action to address infringement, each Party shall prescribe in its law conditions for Internet Service Providers to qualify for the limitations described in para. 1(b), (a) With respect to the functions referred to in para.s 2(c) and 2(d), these conditions shall include a requirement for Internet Service Providers to expeditiously remove or disable access to material residing on their networks or systems upon obtaining actual knowledge of the

was apparent, such as through effective notifications of claimed infringement in accordance with clause (ix); and (C) publicly designating a representative to receive such notifications.



copyright infringement or becoming aware of facts or circumstances from which the infringement is apparent, such as through receiving a notice of alleged infringement from the right holder or a person authorized to act on its behalf. (b) An Internet Service Provider that removes or disables access to material in good faith under subpara. (a) shall be exempt from any liability for having done so, provided that it takes reasonable steps in advance or promptly after to notify the person whose material is removed or disabled.

Table 4.1 (cont.)

Prohibition of requiring monitoring



EU–Republic of Korea

CETA

US–Republic of Korea

USMCA

Article 10.66 No general obligation to monitor

Article 20.11(3)

Article 18.10(30)

Article 20.88

The eligibility for the limitations or exceptions referred to in this Article may not be conditioned on the service provider monitoring its service, or affirmatively seeking facts indicating infringing activity.

1.(b)(vii) Eligibility for the limitations in this subpara. may not be conditioned on the service provider monitoring its service, or affirmatively seeking facts indicating infringing activity, except to the extent consistent with such technical measures.

7. Eligibility for the limitations in para. 1 shall not be conditioned on the Internet Service Provider monitoring its service or affirmatively seeking facts indicating infringing activity, except to the extent consistent with the technical measures identified in para. 6(b).

1. The Parties shall not impose a general obligation on providers, when providing the services covered by Articles 10.63 through 10.65, to monitor the information which they transmit or store, nor a general obligation to actively seek facts or circumstances indicating illegal activity

       



The geographical reach and overlap of the elaborate EU and US agreements illustrates the successful exportation and thus a high degree of assimilation among the relevant RTA partners that will arguably contribute to seamless digital markets in which the basic determinants of ISP liability will be the same, providing some of the desired legal certainty sought by businesses. Although the purpose of each RTA may have been to export the domestic system, countries subject to both obligations will likely comply by adopting the ‘higher’ standard, illustrating the de facto cumulative effect of RTA IP obligations that are formulated as minimum standards. Past experience indicates that signatories of these RTAs are likely to include substantively similar provisions in future agreements further extending the hub-and-spoke architecture, and thus reinforcing this trend of a common regulatory response by virtue of RTAs.

Online Exhaustion The question whether exhaustion – the principle that once an IPprotected good has legitimately entered distribution channels, its further distribution (e.g. of a second-hand book) no longer requires the agreement from the original right owner – could also apply to downloaded digital products (e.g. e-books or software) is determinative for whether a potentially significant global market for ‘second-hand’ digital products could exist.

Exhaustion in International Treaties While the concept of exhaustion – or, for that matter, territoriality – is not explicitly mentioned in the Paris or Berne Conventions,210 a number of interpretations have taken the territoriality of IP protection arguably implied in these conventions to indicate a preference for national exhaustion.211 In the TRIPS context, the intense yet inconclusive discussion of pharmaceutical parallel imports led to the exclusion of exhaustion from dispute settlement considerations212 which leaves WTO members free to

210 211

212

See Article 4bis Paris Convention and Article 5 Berne Convention. See Gervais (2012), para. 2.99. See also Gosh and Calboli (2019). Regarding the Paris Convention, see German Federal Supreme Court, Centrafarm and Dirk de Fluiter v. Eli Lilly & Co., 8 IIC 64 (1977) – Tylosin (supporting national exhaustion); and Tokyo High Court, 27 IIC 550 (1996) (rejecting implications for exhaustion). Article 6 TRIPS Agreement.



 .     -                   

establish its own regime for exhaustion without challenge.213 Where TRIPS text does – indirectly – articulate the concept, exhaustion is formulated exclusively with respect to goods, put on the market by, or with the consent of, the right holder.214 As has been described above,215 the discussion on exhaustion during the negotiations of the Internet Treaties led to the agreed statement concerning Articles 6 and 7 of the WCT, affirming that only physical copies were subject to the new distribution right. While this has generally been understood to mean that exhaustion does not apply to digital products downloaded from the Internet (i.e. ‘made available’ rather than ‘distributed’), the purported flexibility of the ‘umbrella solution’ as a background to the making available right would also seem to permit a more nuanced interpretation.216

National Jurisprudence on Online Exhaustion – Filling the Gaps? In the absence of international agreement on these issues, the role of the courts in applying domestic law can be significant in determining the scope of exhaustion in practice. The increasing use of electronic channels for the purchase of digital products raises the question whether legal distinctions associated with the different modes of online and offline transfer should continue to apply where they create counter-intuitive distinctions between otherwise – from the consumer perspective – identical products.217 In the words of the EU Advocate General Spuznar: The digitisation of content and the development of the new means of supplying that content to users made possible by the internet have upset the balance that existed in the analogue environment between, on the one hand, the interests of copyright holders and, on the other hand, the interests of users of the works. The rule of the exhaustion of the right of 213

214

215 216 217

See Doha Declaration on TRIPS and Public Health WT/MIN(01)/DEC/W/2 of 14 November 2001, para. 5(d): ‘The effect of the provisions in the TRIPS Agreement that are relevant to the exhaustion of IPRs is to leave each Member free to establish its own regime for such exhaustion without challenge, subject to the MFN and national treatment provisions of Articles 3 and 4.’ See footnote 13 to Article 51 TRIPS, excluding from the obligation of border measures ‘imports of goods put on the market in another country by or with the consent of the right holder. This formulation tracks Article 6.5 of the Treaty on Intellectual Property in Respect of Integrated Circuits (1989) which has been incorporated into the TRIPS Agreement by virtue of Article 35 TRIPS Agreement’. See ‘The WIPO Internet Treaties: An Issue-Driven International Solution’ above. Ibid. For a more elaborate discussion and background, see Taubman (2020).

       



distribution is one of the instruments that help to maintain that balance. The question is whether the balancing of the interests involved also requires the application of that rule in the case of the supply of works by downloading.218

In the landmark case UsedSoft,219 the Court of Justice of the European Union decided in the affirmative that copies of software that had been legitimately acquired and downloaded could be resold if the transfer had been definitive and no residual copy remained with the seller. While the decision was taken in the specific context of the EU Software Directive,220 and hence much of its appreciation centred around its characterization of software licensing as tantamount to a ‘first sale’ under Article 4(2) of the Software Directive,221 the Court expressly rejected the Commission’s argument drawn from the WCT background above and confirmed: In those circumstances, it must be considered that the exhaustion of the distribution right under Article 4(2) of Directive 2009/24 concerns both tangible and intangible copies of a computer program, and hence also copies of programs which, on the occasion of their first sale, have been downloaded from the internet onto the first acquirer’s computer.222

Subsequent European jurisprudence has been reluctant to apply this approach outside the scope of the Software Directive and, in the recent Tom Kabinet223 case relating to second-hand e-books, expressly rejected the assimilation of tangible and intangible copies under the Copyright Directive,224 citing – inter alia – its purpose as implementing Articles 6(1) and 8 WCT and referring the agreed statement in that regard. Although this appears as a reversal to the traditional view that exhaustion is limited to the distribution of tangible copies only, the tension of this approach with the economic reality is apparent in the Advocate General’s opinion in this case: The foregoing considerations lead me to conclude that arguments, of both a legal and a teleological nature, are in favour of recognition of the rule of

218 219

220

221 222 223 224

Tom Kabinet, C‑263/18, Opinion of Advocate General Szpunar, para. 79. UsedSoft, C-128/11, 3 July 2012, available at curia.europa.eu/juris/liste.jsf?num=C-128/ 11. Directive 2009/24/EC of the European Parliament and of the Council of 23 April 2009 on the legal protection of computer programs. See Hilty et al. (2013). UsedSoft, C-128/11, 3 July 2012, para. 59. Tom Kabinet, C‑263/18, 19 December 2019. Tom Kabinet, C‑263/18, para. 56.



 .     -                    exhaustion of the distribution right with respect to works supplied by downloading for permanent use. In particular, the permanent possession by the user of a copy of such a work shows the similarity of that mode of supply with the distribution of tangible copies. However, I am of the view that, as EU law now stands, the arguments to the contrary should prevail. These are, in particular, the arguments [. . .] concerning the EU legislature’s clear intention that downloading should be covered by the right of communication to the public, the limitation of the distribution right to acts of transfer of ownership of a copy, and the right of reproduction.225

A similar tension between the limits of existing legislation and the realities of digital trade is apparent in the US jurisprudence related to this question in the cases ReDigi and ClearCorrect. In ReDigi226 the 2nd Circuit Court of Appeals in New York rejected applying exhaustion to lawfully purchased iTunes tracks that were transferred to the purchasing customer while ensuring deletion from the seller’s network. It argued that uploading the music files on ReDigi’s server ‘inevitably involved’ the creation of new unauthorized phonorecords by reproduction, which are excluded from the scope of exhaustion, and rejected a fair use defence in light of the negative impact the use would have on the music market. The ClearCorrect227 case did not concern exhaustion as such, but rather whether the International Trade Court (ITC) had jurisdiction to bar the import of digital datasets related to teeth aligning models.228 The Federal Circuit found that 19 US Code § 1337(a) – relating to importation of ‘articles’ – limited the ITCs jurisdiction to ‘material things’ and did not include the ability to bar digital imports. In a dissenting opinion, Judge Newman argued for a broad interpretation of ‘articles’ to include intangible products, citing several US statutes and administrative statements that had already assimilated tangible and intangible articles and goods.229

Prospects for a Common Regulatory Response This national jurisprudence highlighted above illustrates the increasing tension between the traditional approach to exhaustion as applying only 225 226 227

228

229

Tom Kabinet, C‑263/18, Opinion of Advocate General Szpunar, para. 79. Capitol Records, LLC et al. v. ReDigi Inc., No. 16-2321 (2nd Cir. 2018). ClearCorrect Operating, LLC v. ITC, No. 2014-1527, Slip Op. at 3 (Fed. Cir. November 10, 2015). Discussed in Taubman, Antony, ‘The Shifting Contours of Trade in Knowledge: The New “Trade-Related Aspects” of Intellectual Property’, Chapter 2 in this volume (‘Beyond Trade in Goods?’, p. 62). Ibid., at 1304 et seq.

       



to tangible products and the reality of pervasive digitization. While sympathetic to considering the new business models that challenge traditional exhaustion, the courts consider themselves constrained by existing statutes that enshrine the traditional distinction between tangible and intangible products under copyright. In their efforts to fill existing gaps by creative solutions, no common approach has yet emerged around which initiatives for an international regulatory response could coagulate. Characteristically, in the European context, many of the arguments resisting a relaxation of exhaustion centre around the interpretation of the WCT’s ‘right of making available’ as a communication to the public, and hence the exclusion of downloaded products from exhaustion under the guidance of the agreed statement to Article 6 WCT. As highlighted above in the section the Internet Treaties, WCT signatories that have chosen the alternative means of implementing the ‘right of making available’ as a form of the distribution right230 should feel less constrained by such considerations to apply exhaustion in the digital context. Considerations of compelling WCT interpretation in this regard are thus notably absent from the US jurisprudence cited above. Any recalibration of the balance of interests between right holders and users with regard to digital products will need to be addressed in light of economic realities, taking into account the purpose and objective of the tool of exhaustion. In digital situations that emulate offline exhaustion – such as a permanent transfer of a digital product with no residual copy, as elaborated in UsedSoft – court opinions and policy considerations have shown sympathy for an application of the same principle. It is recognized that in the absence of accessible evidentiary tools to verify such conditions on a large scale as would be required for, say, the monitoring of a particular e-book copy, the application of online exhaustion will need further consideration. However, the new technologies that have upset the traditional balance in the analogue world may yet be instrumental in establishing a new balance. If technologies like blockchain could create a reliable technical standard for evidence for how many copies continue to exist of a particular digital product, this may help address the evidentiary side of a concept of online exhaustion and help spur a common regulatory response in this area.

230

See the references to the US implementation above at nn. 40 and 41.



 .     -                   

New Issues – Rapidly Approaching: Data Mining and Artificial Intelligence Rapid progress in technology and its widespread use means that the next issues requiring a – preferably common – regulatory response are already approaching. Among them are the questions of how to address the increasing commercial and ethical relevance of ‘big data’, and how to address the occurrence of artificial intelligence (AI) in the traditionally human-centric system of intellectual property protection.

Big Data and Emerging Data Mining Exceptions Cutting across a number of different disciplines, the policy approaches to harnessing the welfare potential of using ‘big data’ collection and use have often been characterized as being similar to that of IP – namely striking the right balance between incentivizing the creation (i.e. collection) of data while ensuring sufficient access and benefit for users from the general public. Seeking to contribute to the latter, the practice of text and data mining – i.e. provisions intended to permit the analysis of large volumes of digital text and data to identify patterns, trends and correlations – has been recognized in legislative initiatives in Europe,231 and under the fair use principle in the United States.232 The, as yet divergent and incomplete, regulatory approaches in this area reflect that the application of IP protection or exceptions to large collections of data have not yet been fully settled. Challenges lie not only in different approaches to traditional database approaches – such as a sui generis database right in the European Union233 – but also stem from the evolving nature of data collections that resist categorization in traditional IP right categories.234 Artificial Intelligence in a Human-Centric IP System AI tools and their outputs pose a number of conceptual challenges to the IP system which has traditionally rewarded inventive and creative 231

232

233 234

See Articles 3 and 4 of the DSM Directive and the background to criticism of its limited scope in EU Parliament (20018a) and EU Parliament (20018b), and Benhamou (2020). See Authors Guild v. HathiTrust, 755 F.3d 87 (2nd Cir. 2014), White v. West (SDNY 2014), Fox v. TVEyes (SDNY 2014), Authors Guild v. Google, 770 F.Supp.2d 666 (SDNY 2011). See Benhamou (2020) and Gervais (2019). See Gervais (2019) ‘Beyond the protection of software used to collect and process Big Data corpora, copyright’s traditional role is challenged by the relatively unstructured nature of the non-relational (noSQL) databases typical of Big Data corpora.’

       



activity associated with humans. AI tools, such as machine learning, big data analysis and evolutionary algorithms pose questions as to IP ownership, standards for IP protection criteria such as ‘obviousness’235 and ‘originality’, as well as the protectability of AI outputs. Responses to these challenges have so far triggered clarifications in patent examination guidelines236 and isolated national jurisprudence applying existing concepts,237 but broader conceptual engagement with AI at the policy level remains at an early stage. At the international level, at the time of writing, WIPO launched a public consultation process on AI and IP policy to identify the most pressing questions likely to face IP policymakers in this area.238 While responses to the first issues paper239 suggest that considerable clarification is still necessary to properly frame the relevant questions, international coordination in this area could benefit from considerable WIPO groundwork on AI that reaches back as far as the early 1990s.240

Conclusion The mechanisms that influence the creation of international regulatory responses to emerging IP issues have undergone significant change since the advent of the digital era. The primacy of state sovereignty in international affairs in the nineteenth and twentieth centuries, coupled with the unique demands of rapid internationalization of arts and technology at the turn of the century, led to a dominance of international organizations as the forum for developing and agreeing international regulatory responses in the area of IP. The favouring of consensus decisions and

235 236

237

238 239 240

Abbott (2018). See, for example, Singapore’s April 2019 AI-related revision of Examination Guidelines for Patent Applications. See patent applications in the name of an IT tool ‘DABUS’ filed in multiple jurisdictions by The Artificial Inventor Project (artificialinventor.com). In November 2019 the EPO rejected two such applications (EP 18275 163 and EP 18275174) on the basis of the requirement that an inventor designated in the application has to be a human being, not a machine (see www.epo.org/news-issues/news/2020/20200128.html). See WIPO press release PR/2019/843 of 13 December 2019. See MPI (2020) for an elaboration of the conceptual challenges. See WIPO Worldwide Symposium on the Intellectual Property Aspects of Artificial Intelligence held at Stanford University 25–27 March 1991, available at www.wipo.int/ edocs/pubdocs/en/wipo_pub_698.pdf.



 .     -                   

the iterative process of revising the treaties created a situation in which technical regulatory responses developed and experienced at the national level could be directly translated into binding international rule-making. Coinciding with the advent of the digital era, this scenario has given way to a more varied landscape where a variety of tools and legal instruments embody regulatory responses to digital challenges. While TRIPS, the WCT and the WPPT are examples of the traditional model of multilaterally negotiated treaties in existing institutions, the adoption of non-binding instruments – such as the Joint Recommendations – shows the limits of consensus and momentum in this model. International nonstate solutions such as ICANN and the UDRP emerge as separate, technically determined solutions to specific IP issues. Proliferating FTAs emerge as a new platform to formulate and agree IP-related regulatory responses that can be used to project the national solutions of one FTA-partner but can also serve to give legally binding status to internationally agreed non-binding recommendations. Today, these different modes of international regulation co-exist and interact, reflecting ‘the reality of a multi-speed and multi-tiered world in which multilateralism, while being the highest expression of inclusiveness and legitimacy, is nevertheless the slowest solution’.241 This chapter shows how these diverse approaches interact in recent IP regulatory responses to emerging digital issues that are particularly relevant for digital business models. In the relatively mature area of ISP liability, global regulatory responses show broad agreement in principle that limitation of liability is essential for the viability of Internet platforms and the important functions they perform in the global digital marketplace. While the methodology on conditions and sanctions of the various ‘safe harbour’ models still varies, broad agreement in principle has meant that the concept has been included in bilateral and regional trade agreements. Whether this means that this issue is closer to becoming an international rule will have to be assessed in light of the recent policy changes at the domestic level – notably under the European Digital Single Market Directive (DSM Directive or DSMD). Still further behind are the emerging issues of online exhaustion, data mining and IP-related questions of artificial intelligence. As regards the former, domestic jurisprudence in different countries is still struggling to find a consistent

241

See Gurry acceptance speech 2014 (n. 143).

        



understanding of under what evidentiary circumstances the offline concept can be applied to digital works. With regard to the latter two, a conceptually consistent policy approach is yet to unite the hitherto diverse regulatory reactions ranging from IP office guidelines on IP ownership to data mining exceptions in national law.

Bibliography Abbott, R., ‘Everything is Obvious’ (2018), UCLA Law Review 66, 2, available at https://ssrn.com/abstract=3056915 or http://dx.doi.org/10.2139/ssrn.3056915. Agarwal, Devika and Agarwal, Radhika, ‘Delhi High Court Upholds Safe Harbour for Intermediaries in Copyright Disputes’ (2017), Journal of Intellectual Property Law & Practice 12, 5: 366–368, available at https://doi.org/10 .1093/jiplp/jpx040. Benhamou, Yaniv, ‘Licensing Big Data: A Holistic Analysis Based on a Three-Step Approach’ (2020), available at www.researchgate.net/publication/339140374_ Licensing_Big_Data_a_holistic_analysis_based_on_a_three-step_approach. Bhagwati, Jagdish, ‘US Trade Policy: The Infatuation with Free Trade Agreements’, in J Bhagwati and A Krueger, The Dangerous Drift to Preferential Trade Agreements, Washington, DC: AEI Press, 1995. Blomqvist, Jorgen, A Primer on International Copyright and Related Rights, Cheltenham: Edward Elgar Publishing, 2014. Dinwoodie, Graeme B., ‘(National) Trademark Laws and the (Non-National) Domain Name System’ (2000), University of Pennsylvania. Journal of International Law 495, 21, available at https://scholarship.law.upenn.edu/ jil/vol21/iss3/2. EU Parliament (2018), DG for Internal Policies, (Nordemann) Liability of Online Service Providers for Copyrighted Content – Regulatory Action Needed? 2018 (PE 614.207). EU Parliament (2018a), DG for Internal Policies, (Geiger et al.) The Exception for Text and Data Mining (TDM) in the Proposed Directive on Copyright in the Digital Single Market – Legal Aspects 2018 (PE 604.941). EU Parliament (2018b), DG for Internal Policies, (Rosati) The Exception for Text and Data Mining (TDM) in the Proposed Directive on Copyright in the Digital Single Market – Technical Aspects 2018 (PE 604.942). Ficsor, Mihály, The Law of Copyright and the Internet: the 1996 WIPO Treaties, Their Interpretation and Implementation, Oxford: Oxford University Press, 2002. Ficsor, Mihály, ‘Copyright In The Digital Environment: The WIPO Copyright Treaty (WCT) And The WIPO Performances And Phonograms Treaty (WPPT) in WIPO/CR/KRT/05/7’, February 2005, available at www.wipo .int › mdocs › arab › wipo_cr_krt_05 › wipo_cr_krt_05_7.



 .     -                   

Ficsor, Mihály, ‘Only once more — and then Merry Christmas and Happy New Year to everybody, including Professor Geist and his devoted followers: the 1996 WIPO Diplomatic Conference, the WIPO Treaties and the balance of interests’, Barry Sookman, 23 December 2009, available at www.barrysookman .com/2009/12/23/only-once-more-and-then-marry-christmas-and-happynew-year-to-everybody-including-professor-geist-and-his-devoted-followersthe-1996-wipo-diplomatic-conference-the-wipo-treaties-and-the-balanc/. Ficsor, Mihály, ‘The WIPO Internet Treaties and Copyright in the “Cloud”’, paper presented at the 2012 ALAI Congress in Kyoto. Geist, Michael A., ‘Fair.Com?: An Examination of the Allegations of Systemic Unfairness in the ICANN UDRP’, August 2001, available at https://ssrn .com/abstract=280630. Geist, Michael A., ‘Fundamentally Fair.com? An Update on Bias Allegations and the ICANN UDRP’, 2002, available at http://aix1.uottawa.ca/~geist/ fairupdate.pdf. Geist, Michael A., ‘The Case for Flexibility in Implementing the WIPO Internet Treaties: An Examination of the Anti-Circumvention Requirements’, 2010, available at www.irwinlaw.com/sites/default/files/attached/CCDA%2008% 20Geist.pdf. Gervais, Daniel, The TRIPS Agreement – Drafting History and Analysis, 4th ed., London: Sweet & Maxwell, 2012. Gervais, Daniel, ‘Exploring the Interfaces Between Big Data and Intellectual Property Law’ (2019), Journal of Intellectual Property, Information Technology and Electronic Commerce Law, 10, 22: para. 1. Gosha, Shuba and Calboli, Irene, Exhausting Intellectual Property Rights – A comparative Law and Policy Analysis, Cambridge: Cambridge University Press, 2019. Haas, Marie-Emmanuelle, The origin of the UDRP: NSI’s 1995 domain name dispute policy, (domainesinfo.fr 2009) available at www.me-haas.eu/wp-con tent/uploads/2014/03/The-origin-of-the-UDRP.pdf Helfer, Laurence R. and Dinwoodie, Graeme B., ‘Designing Non-National Systems: The Case of the Uniform Domain Name Dispute Resolution Policy’ (2001), William and Mary Law Review 43: 141, available at https://scholarship.law .wm.edu/wmlr/vol43/iss1/7. Hilty, Reto, Köklü, Kaya and Hafenbrädl, Fabian, ‘Software Agreements: Stocktaking and Outlook Lessons from the UsedSoft v. Oracle Case from a Comparative Law Perspective’ (2013), International Review of Intellectual Property and Competition Law 44: 263–292. Husovec, Martin, ‘Accountable, Not Liable: Injunctions Against Intermediaries’, 2 May 2016, TILEC Discussion Paper No. 2016-012, available at https://ssrn .com/abstract=2773768 or http://dx.doi.org/10.2139/ssrn.2773768.

        



Inta White Paper, The Intersection of Trademarks and Domain Names, International Trademark Association, 18 August 1997, available at www.ntia.doc.gov/legacy/ ntiahome/domainname/not-emailed/INTA-whitepaper.htm. Isenberg, Doug, These Countries Have Adopted the UDRP, 1 June 2017, CircleID, available at www.circleid.com/posts/20170601_these_countries_ have_adopted_the_udrp/. Komaitis, Konstantinos, The Current State of Domain Regulation: Domain Names as Second-Class Citizens in a Mark-Dominated World, New York: Routledge, 2010. Kumar, Aakanksha, ‘Internet Intermediary (ISP) Liability for Contributory Copyright Infringement in USA and India: Lack of Uniformity as a Trade Barrier’ (2014), Journal of Intellectual Property Rights 19: 272–281. Kur, Annette, ‘Not Prior in Time, But Superior in Right: How Trademark Registrations Can be Affected by Third Party Interests in a Sign’ (2013), International Review of Intellectual Property and Competition Law 44, 7: 790–814, available at https://ssrn.com/abstract=2253888. Kwakwa, Edward and Talbott, Autumn, ‘The Influence of the World Intellectual Property Organization on the European Union’, in Ramses Wessel and Steven Blockmans (eds), Between Autonomy and Dependence: The EU Legal Order under the Influence of International Organizations, The Hague: Asser Press/Springer, 2013. LaFrance, Mary, Streaming of Music and Audiovisual Works, Chapter 19 in this volume, 2020. von Lewinski, Silke, ‘Article 14 WPPT – Right of Making Available of Phonograms’, in Jorge Reinbothe and Silke von Lewinski, The WIPO Treaties on Copyright: A Commentary on the WCT, the WPPT and the BTAP, Oxford: Oxford University Press, 2015. MPI, Comments of the Max Planck Institute for Innovation and Competition of 11 February 2020 on the Draft Issues Paper of the World Intellectual Property Organization on Intellectual Property Policy and Artificial Intelligence 2020, available at www.ip.mpg.de/fileadmin/ipmpg/content/stel lungnahmen/2020-02-11_WIPO_AI_Draft_Issue_Paper__Comments_Max_ Planck.pdf. Mulholland, Caitlin, Secondary liability of service providers in Brazil: The effect of the Civil Rights Framework, (responsabildadecivil.org) 29 March 2018, available at www.responsabilidadecivil.org/single-post/2018/03/29/ Secondary-liability-of-service-providers-in-Brazil-the-effect-of-the-CivilRights-Framework. Otten, Adrian, ‘The TRIPS Negotiations: An Overview’, in Jayashree Watal and Antony Taubman (eds), The making of the TRIPS Agreement, Geneva: WTO, 2015, pp. 55–79.



 .     -                   

Petillion, Flip and Janssen, Jan, Competing for the Internet: ICANN Gate – An Analysis and Plea for Judicial Review through Arbitration, Alphen aan den Rijn, Netherlands: Kluwer, 2017. Ricketson, S., The Paris Convention for the Protection of Industrial Property – A Commentary, Oxford: Oxford University Press, 2015. Ricketson, S. and Ginsburg, J., International Copyright and Neighbouring Rights – The Berne Convention and Beyond, 2nd ed., Oxford: Oxford University Press, 2006. Rotaru, Vasile, Accountable Algorithmic Copyright Enforcement, PIR 2017/2018, available at www.academia.edu/37432171/ACCOUNTABLE_ALGORITHMIC _COPYRIGHT_ENFORCEMENT. Saha, Ishaan and Saha, Oieshi, ‘Safe Harbour Law In India: Raising the Red Flag in the Wake of Myspace v. Super Cassettes’ (2018), Journal of Intellectual Property Law & Practice 13, 5: 393–399, available at https://doi.org/10 .1093/jiplp/jpx192. Schlesinger, Michael, ‘The Right of Making Available and its Implementation in National Law and Case Law’, in Hugh Hansen (ed.), Intellectual Property Law and Policy, Vol. 11, Oxford: Hart, 2010. Schwemer, Sebastian and Schovsbo, Jens, ‘What is Left of User Rights? – Algorithmic Copyright Enforcement and Free Speech in the Light of the Article 17 Regime’, forthcoming in Paul Torremans (ed.), Intellectual Property Law and Human Rights, 4th ed., Alphen aan den Rijn, The Netherlands: Wolter Kluwers, 2020. Senftleben, Martin, ‘Institutionalized Algorithmic Enforcement – The Pros and Cons of the EU Approach to UGC Platform Liability’ (2020), Florida International University Law Review 14, available at http://ssrn.com/ abstract=3565175. Sethia, Aradhya, ‘The Troubled Waters of Copyright Safe Harbours in India’ (2017), Journal of Intellectual Property Law & Practice 12, 5: 398–407, available at https://doi.org/10.1093/jiplp/jpx017. Seuba, Xavier, ‘Substantive and Jurisdictional Challenges Arising from Bilateralism in Intellectual Property’ in CEIPI – ICTSD, What Institutional Environment for the Development and Enforcement of IP Law?, November 2015, CEIPI – ICTSD Series: Global Perspectives and Challenges for the Intellectual Property System; Issue No. 1, pp. 57–68. Stutz, Michael, ‘NSI Tweaks Domain Name Dispute Rules’, wired.com 2 December 1998, available at www.wired.com/1998/02/nsi-tweaks-domain-disputerules/. Taubman, Antony, The Lasagne Effect: Bilateral, Regional and Multilateral Standards, and the Settlement of Disputes Over Trade and Intellectual Property, Society of International Economic Law Inaugural Conference, Geneva, July 2008.

        



Taubman, Antony, ‘Discontent Industries? Creative Works and International Trade Law: Making Sense of “Analogue” IP Rules in a Digital Age’, 9 November 2019, available at https://ssrn.com/abstract=3566680. Taubman, Antony, ‘Digital Disruption: Reshaping Markets for IP’, forthcoming in Robert D. Anderson, Nuno Pires de Carvalho and Antony Taubman, Competition Policy and Intellectual Property in Today’s Global Economy, Cambridge: Cambridge University Press, 2020. US Copyright Office, The Making Available Right in the United States – A Report of the Register of Copyrights, 22 February 2016. US Green Paper, National Telecommunications and Information Administration (NTIA), A Proposal to Improve Technical Management of Internet Names and Addresses – Discussion Draft 1/30/1998, available at www.ntia.doc.gov/ legacy/ntiahome/domainname/dnsdrft.htm. US White Paper, US Department of Commerce, Statement of Policy on the Management of Internet Names and Addresses, 5 June 1998, available at www.ntia.doc.gov/federal-register-notice/1998/statement-policy-manage ment-internet-names-and-addresses. Valdés, R. and McCann, M., ‘Intellectual Property Provisions in Regional Trade Agreements: Revision and Update’, WTO Staff Working Paper ERSD-201414, Geneva: World Trade Organization, 2014. Valdés, R. and Runyowa, T., ‘Intellectual Property Provisions in Regional Trade Agreements’, WTO Staff Working Paper ERSD-2012-21, Geneva: World Trade Organization, 2012. Wager, Hannu, ‘Copyright: A Nordic Perspective’, in Jayashree Watal and Antony Taubman (eds), The making of the TRIPS Agreement, Geneva: WTO, 2015, pp. 321–340. Watal, J. and Taubman, A. (eds) (2015), ‘The Making of the TRIPS Agreement: Personal Insights from the Uruguay Round Negotiations’, Geneva: WTO, 2015, available at https://doi.org/10.30875/00344762-en. De Werra, Jacques, ‘Alternative Dispute Resolution Mechanisms for Solving Trademark Disputes (Mediation, UDRP, Arbitration)’, in Irene Calboli and Jacques de Werra (eds), The Law and Practice of Trademark Transactions: A Global and Local Outlook, Cheltenham: Edward Elgar Publishing Ltd, 2016, pp. 293–322, available at https://archive-ouverte .unige.ch/unige:82027. WIPO, Final Report of the WIPO Internet Domain Name Process, 30 April 1999, available at www.wipo.int/amc/en/processes/process1/report/finalreport.html. WIPO, Intellectual Property Handbook, 2nd ed., Geneva: World Intellectual Property Organization, 2004. Zingales, N., ‘The Brazilian Approach to Internet Intermediary Liability: Blueprint for a Global Regime?’ (2015), Internet Policy Review 4, 4: DOI: 10.14763/ 2015.4.395.

PART II Measuring Trade in Knowledge

5 Measuring International Intellectual Property Transactions in a Globalized World: Current Challenges and Possible Improvements                                  * Abstract Trade in IP products (IPP) has grown rapidly over the past decades and at an accelerating pace in recent years. Advances in information and communications technologies (ICTs), and the possibilities offered by digitization and new business models, have widened opportunities for such trade, shaping production networks in goods and services and global value chains. This paper describes the importance of such trade. It depicts current concepts and methods used to compile official statistics in a global value chain environment and to describe characteristics of traders in IP products. In considering what we ca learn from cross-border knowledge flows from IMF Balance of payments data and other statistical frameworks, it concludes with an outline for a way forward to improve the information base.

Introduction Intellectual property (IP) is defined as ‘the rights relating to: literary, artistic and scientific works, performances of performing artists, phonograms, and broadcasts, inventions in all fields of human endeavour, scientific discoveries, industrial designs, trademarks, service marks, and commercial names and designations, protection against unfair competition, and all other rights resulting from intellectual activity in the industrial, scientific, literary or artistic fields’.1 It is described as ‘one of the most readily tradable properties in the digital marketplace’.2 Thus, IP is an essential component of the production of goods and

* Thanks go to Florian Eberth and Barbara d’Andrea for providing statistics and text on high-tech goods and IP trade. 1 See www.wipo.int/treaties/en/text.jsp?file_id=283854#P50_1504. 2 See www.businessdictionary.com/definition/intellectual-property.html.





      

services, particularly for those that are innovative or creative. It is traded internationally: i. embedded in the value of traded goods and services ii. through international licensing transactions, or iii. through the sales/transfer of ownership in IP rights (including the results of Research and Development – R&D, assignment of IP portfolios through acquisition, or rights on computer programs, cinematographic works or sound recordings). Trade in IP products (IPP)3 has grown rapidly over the past decades and at an accelerating pace in recent years. Advances in information and communications technologies (ICTs), and the possibilities offered by digitization and new business models, have widened opportunities for such trade, shaping production networks in goods and services and global value chains. Statistically, the Manual on Statistics of International Trade in Services 2010 (MSITS2010) as well as the revised Balance of Payments Manual, 6th edition (BPM6) and accompanying compilation guidance have clarified the concepts of intellectual property in statistical frameworks. They present a classification system of related receipts and payments, depending on how IP is transacted internationally (embedded in products, licensing, sales or transfer of ownership of IP rights), and at what level of detail (see Box 5.1). However, globalization, digitalization and tax optimization have contributed to blur the way transactions are recorded and compiled into relevant statistics for policymakers, research and for businesses themselves. MSITS2010 catalogues research and development services as well as transactions in computer software, audiovisual and cultural products. Compilation guides of the BPM6 and MSITS2010, as well as the Guide to Measuring Global Production of the Conference of European Statisticians, which deals with national accounting aspects, provide further guidance as to how compilers could approach some of the challenging measurement issues arising from the globalization of the economy. But activities of multinational enterprises (MNEs) make it more complex to value international transactions and assess related transfers of ownership.

3

These products include research and development, computer software and information services, audiovisual and other cultural products, franchises and trademarks.

     



The shortage of data makes it difficult to answer important policy questions. For instance, to what extent are less developed economies benefiting from the expansion of trade and investment in technology? What are the characteristics of firms which are engaging in IP international transactions? What is the role of small and medium-sized enterprises in IP trade? This chapter will review what developments are needed for compiling relevant statistics to respond to these important questions. The first part of the chapter will present some general trends in international transactions pertaining to IP. The second part will review the concepts and definitions related to IP which are currently available in trade in goods and services statistics, and report on the current data availability. The third part will illustrate some of the measurement challenges that compilers face to develop such data. Linkages between balance-of-payments IP-related transactions, foreign direct investment (FDI) and foreign affiliates statistics (FATS) will be shown. The fourth part will explore possibilities offered by statistics on trade by enterprise characteristics in relation to trade in IP. The final part will sketch a possible way forward for the collection and compilation of data (e.g. improve or create new surveys, finding synergies) to provide a better understanding of the economics of trade, global value chains and IP.

Importance of Intellectual Property in International Trade Different types of statistics can be used to assess the importance of IP in international trade. These cover customs statistics (trade in high-tech and branded goods, and specific goods such as recorded optical media and printed publications) and the balance-of-payments (BPM6) for services (including charges for the use of intellectual property and research and development services). For intra-firm transactions, FDI statistics and FATS are useful frames. Quantitative indicators such as patent applications, grouped by family, filed abroad by country of origin, valued by citations, patent applications with at least one applicant from abroad, patent assignments across borders, etc. are a complementary data source. In this chapter we focus on transactions recorded as goods and services in the balance of payments. Since 1995, the value of exports in high-tech goods, which includes a very large share of IP, has increased significantly. For developed countries, it increased by nearly 80 per cent, while their share of this global trade decreased from 72 to 51 per cent. In contrast, the developing



      

Table 5.1 Exports of high-tech goods 1995, 2000, 2010, 2015–20174 (Billion US dollars and percentage)

Reporter group World Developed economies Developing economies & CIS

Value ($bn) High-tech goods 1995 2000 2010 586 966 1286 424 634 689 162 331 597

2015 1480 741 739

2016 1437 744 693

2017 1480 762 718

Reporter group World Developed economies Developing economies & CIS

Share (%) High-tech goods 1995 2000 2010 100 100 100 72 66 54 28 34 46

2015 100 50 50

2016 100 52 48

2017 100 51 49

Source: WTO, UN Comtrade data.

economies’ share increased by 75 per cent, its value growing more than four times to reach USD 718 billion in 2017 (see Table 5.1). Trade in charges for the use of IP, which includes licence fees for the use of IP, has increased more than twice as fast as exports of high-tech goods. The increase in such trade is more than four-fold since 1995, which is 1.5 times faster than the overall increase of trade in commercial services. It accounted for USD 398 billion in 2017 (Figure 5.1). Also, Figure 5.2 shows the increase of transactions relating to R&D services, computer and information and audio-visual services, which was even stronger, in particular since 2005 (no comparable statistics are available for previous years for these services categories). Transactions referring to the use of IP are highly concentrated. As seen from Figure 5.3, in 2016, the biggest traders alone, the US and the EU (28), held a share of more than 76 per cent. Almost half of the European Union’s receipts of charges and licensing fees originated from within the EU, and receipts from the United States accounted for an additional 20 per cent. Similarly, in the case of the United States, around 40 per cent of receipts were arising from exports to EU member countries, in particular Ireland and the United Kingdom (at this time a member of the EU), as well as from exports to Switzerland and Canada. Data on payments also reflect 4

For high-tech goods, the OECD-Definition in document OCDE/GD(97)216) is applied.

     



400 300 200 100 0

Charges for the use of intellectual property n.i.e.

Figure 5.1 World international transactions relating to charges for the use of intellectual property, n.i.e., 1995–2017 (billion US dollars) Source: ITC, UNCTAD, WTO, 2018. 600000

CS, R&D, AV

500000 400000 300000

CUIP

200000 100000 0

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

Figure 5.2 World international transactions relating to charges for the use of intellectual property n.i.e. (CUIP), computer services (CS), research & development (R&D) and audio-visual services (AV), 2005–2017 (Source: IMF) Source: ITC, UNCTAD, WTO, 2018.

these trends. US data reflect that a large share of IP international transactions takes place within multinational enterprises.

Concepts Pertaining to IP-Related International Transactions and Current Data Availability Relevant international statistical guidelines strove to cover innovation and related IP flows in one way or the other, whether in economic statistics (System of National Accounts 2008, BPM6, Manuals on Statistics of International Trade in Services, OECD Handbook on Deriving Capital Measures of IPPs, OECD Handbook on Economic Globalization Indicators), specialized statistics on innovation and



      

European Union (28)

United States

Switzerland 4.9%

Switzerland 8.8%

Figure 5.3 European Union (28) and the United States – charges for the use of intellectual property, n.i.e., receipts by main partner, 20165 Source: WTO World Trade Statistical Review (WTSR 2018), data are sourced from Eurostat and OECD.

research and development (Frascati Manual, Oslo Manual), as well as quantitative data on patents, trademarks, etc. IP products can be accessed or produced in various ways: within the firm, or through a diffusion process such as accessing open source IP or knowledge, IP acquired from other economic entities (outsourcing of R&D, licensing for the use of IP), or through the purchase of new innovative products (materials, software etc.). We will focus in this chapter on the market international transactions involving IP. In economic statistics, IP is generally described as taking the following forms: • Intellectual property-related products (IPPs), which can take the form of • produced assets: ▪ computer software and databases, audio-visual products (music, movies, television and radio programmes) ▪ outcomes of research & development (R&D) ▪ exploration rights (in general no international related transactions as the client will normally have an entity resident in same economy as owner of rights (in principle government entity) and • non-produced assets (use of trademark). • Exchanges of tangible products: • intellectual property embedded in goods transactions, e.g. cultural or creative goods (mass-produced), • R&D results embedded in goods (e.g. industrial processes), • originals/blueprints exchanged on physical media. 5

See World Trade Statistical Review, WTO, 2018, p. 54.

     



• Exchanges of intangible products (transactions in intangibles): • downloaded products which have a goods equivalent, the difference being the media through which it is delivered/transmitted to clients (i.e. electronic transmission rather than CDs, DVDs or other physical media). Licences for the use of intellectual property, generally transactions • between the owner of originals and producers, distributors which replicate the original, and trademark payments (including franchising type transactions when not possible to separate IP component from other types of transactions, e.g. services). • Transfers of blueprints etc. to producers (most probably also in the form of a licence) or directly to households such as for 3D printing. How has this translated in international guidelines for the recording of international transactions, and what is the underlying data availability? Up to 2008, the concepts and definitions of the Balance of Payments, 5th edition (BPM5) were used for compiling statistics on international transactions. Under the BPM5, sales/transfer of IP products (whether results of R&D, or computer programs, cinematographic works or sound recordings) were recorded indistinguishably as capital account transactions under the item acquisition/disposal of non-produced, non-financial assets. But for other IP-related transactions, as shown in Table 5.2, data availability greatly improved during the late 1990s–early 2000s. For example, by 2010 there were 138 economies reporting receipts for the item royalties and licence fees and 155 for payments. Reporting for royalties and licence fees even surpassed related items, such as computer and information services, in particular on the payments side (142 economies, of which around two-thirds reported computer services separately) or audiovisual and related services (approximately 80). Only 69 economies reported data for research and development services and for advertising and market research services. In terms of breakdowns, many agencies developed estimates for franchises and similar rights (approximately 35–40 reporters, mainly EU and/or OECD economies), separately from other royalties and licence fees. Only a handful of economies went a step further in breaking down royalties and licence fees into more detailed subcategories. These included Australia (industrial processes, computer software, music), Canada (trademarks, franchises, patents and industrial design, copyrights and related rights, software and other royalties), New Zealand (software royalties and licence fees) and the United States (trademarks, franchise fees, industrial processes, books, records and tapes,



      

Table 5.2 Reporting of BPM5 intellectual property-related items, 2010

S262 Computer and information services S263 Computer services S264 Information services S266 Royalties and licence fees S891 Franchises and similar rights S268 Other business services S273 Miscellaneous business, professional and technical services S277 Business and management consultancy, public relations services S278 Advertising, market research, etc. S279 Research and development services S280 Architectural, engineering, other technical consults S287 Personal, cultural and recreational services S288 Audiovisual and related services S289 Other personal, cultural and recreational services S895 Personal services, education S896 Personal services, health

Receipts

Payments

133 94 75 138 34 167 107

142 106 94 155 38 174 112

50

66

68 69 68 123 76 73 28 27

73 68 64 133 82 80 28 28

Source: IMF.

broadcasting and recording of live events, general use computer services, industrial engineering services). However, as can be seen, this was far from closing the gap in terms of information needs relating to international IPrelated transactions. In addition, there were many issues pertaining to the delineation between different items. BPM6, which was released in 2008, introduced major conceptual improvements. A consistent complement to BPM6 is MSITS2010. The MSITS2010 provides for more details than BPM6 and introduces more detail for some items. Both these manuals represent a major step forward to respond to information needs. First, originals (i.e. the IP product over which ownership rights can be established, e.g. the master copy of a film, music, software, patented result of R&D) are now considered as produced assets (i.e. the result of a (creative) production process). The results of research and development (R&D) such as patents, copyrights, and industrial processes are now treated as produced assets and included in research and development services (under other business services). Likewise, sales/transfer of IP rights are identified separately under the services that produce them (e.g. computer software originals are

     



identified as a separate category under computer services). The only IPrelated product that remains under the capital account relates to marketing assets such as brand names, trademarks, logos, etc. Second, the breakdown of the item ‘other business services’ is reorganized, enabling the separate identification of research and development services (R&D). This provides more visibility to an important item in terms of international trade, which was previously neglected in international statistical guidelines. Third, the delineation is clarified between goods, services and different types of licences for use for some of the most important IP-related products in international transactions (i.e. R&D, computer and audio-visual services). Box 5.1 summarizes the treatment of intellectual property in BPM6 and MSITS2010 guidelines. The classification of transactions relating to the use of those originals (copies, reproduction or distribution rights) will depend on several factors. Charges for the use of intellectual property in MSITS2010 and BPM6 cover the charges for use of proprietary rights (patents, trademarks, copyrights, industrial processes and designs, franchises, etc.) whether the rights result from research and development, marketing, advertising or other creative or cultural activities (computer software, music, films, books, etc.). This item corresponds to the item royalties and licence fees in BPM5. Charges for the use of intellectual property includes licensing receipts or payment for reproduction and/or distribution of computer software, audiovisual products or other products (i.e. the transactions between the owners of the original/rights and producers who either reproduce the products or embed them in other products). However, it is important to note that the item as defined in BPM6 and MSITS2010 (i.e. charges for the use of intellectual property not included elsewhere) does not cover the value of end-user licences or copies themselves (i.e. the licence component of music, video or software sold at the level of the consumer). Transactions with an end-user licence/copies which are sold on a physical media (such as a music CD, or a movie on a DVD) will either be classified under goods if they are mass-produced products with a licence for perpetual use (in Box 5.1 identified as non-customized products provided on physical media with right to perpetual use) or as services (i.e. depending on the type of product, under computer software, audiovisual services etc.) if the sale entails the subsequent payment(s) of periodic licence fees, for example, for updates. If they are customized products, they will be classified as services. All copies delivered or distributed electronically (downloading, online consumption, streaming) will be classified under the relevant services. These are products of great interest in



      

 .      -- 

Use of intellectual property Franchises and trademarks

Charges for the use of intellectual property n.i.e.

Outcomes of research and development Computer software products; audiovisual and related products

Charges for the use of intellectual property n.i.e.

(a) Customized products of all types (b) Non-customized products, downloaded or otherwise electronically delivered (c) Non-customized products provided on physical media with periodic licence fee (d) Non-customized products provided on physical media with right to perpetual use a

Licence to use excluding reproduction and distributionb Relevant service itemd Relevant service itemd

Sale or purchase of intellectual property ownership rightsa Balance-ofpayments capital account Research and development services

Licence to reproduce and/or distributec Charges for the use of intellectual property n.i.e.

Relevant service itemd

Relevant service itemd

Goods

Covers the case where there is a change of economic ownership of the whole of the intellectual property right in question and the seller no longer has any

     



 . (cont.) rights or obligations associated with the intellectual property. This case also includes second or subsequent outright sales of intellectual property rights. b Covers the case where a specific product is supplied with the right to use the intellectual property embodied in it, but not to copy it for further distribution. The transactions should be classified under the appropriate goods and services items. c Covers the case where authority to reproduce and/or distribute the intellectual property is delegated by its owner. d The relevant item is classified under either computer services or audiovisual and related products, depending on the nature of the content provided. For example, the sale or purchase of a copy of a software package that is mass-produced, and is obtained by an individual to load onto a single computer is covered by a licence to use that excludes reproduction and distribution. If a manufacturer pays for the right to include the software on the computers that it produces, then the payment would be a licence to reproduce and/or distribute (charges for the use of intellectual property provided by the owner of the original). Source: Table III.1 in Manual on Statistics of International Trade in Services, 2010 edition.

the context of trade in digitized products.6 Corresponding international transactions, when they occur, will be reported indistinguishably in the relevant item together with the other transactions likely to fall under these services (e.g. computer services, services provided by firms/individuals for the shooting of a movie, transactions in originals, etc.). The relevant items in the most detailed breakdown provided by MSITS2010 are shown in Table 5.3, together with the number of those reporting each individual item for receipts and payments. Although the number of those reporting in BPM5 and BPM6/MSITS2010 has not drastically changed for the main item of interest here, that is charges for the use of intellectual property not included elsewhere (hereafter n.i.e.) (still in the range of 140 for receipts and 160 for payments), those reporting other main items have increased substantially. Research and development show the highest increase, with 50 per cent more reporters (up to approximately 100). The second-best performer is computer 6

See Erick S. Oh, ‘The Global Digital Content Landscape’, Chapter 10 in this volume.



      

Table 5.3 Reporting of BPM6/MSITS2010 intellectual property-related items, 2015

8 Charges for the use of intellectual property, n.i.e. 8.1 Franchises and trademarks licensing fees 8.2 Licences for the use of outcomes of research and development 8.3 Licences to reproduce and/or distribute computer software 8.4 Licences to reproduce and/or distribute audiovisual and related products 8.4.1 Licences to reproduce and or distribute audiovisual products 8.4.2 Licences to reproduce and/or distribute other products 9 Telecommunications, computer, and information services 9.2 Computer services 9.2.1 Computer software of which: 9.2.1.a Software originals 9.3 Information services 9.3.2 Other information services 10 Other business services 10.1 Research and development services 10.1.1 Work undertaken on a systematic basis to increase the stock of knowledge 10.1.1.1 Provision of customized and noncustomized R&D services 10.1.1.2 Sale of proprietary rights arising from R&D 10.1.1.2.1 Patents 10.1.1.2.2 Copyrights arising from R&D 10.1.1.2.3 Industrial processes and designs 10.1.1.2.4 Other 10.1.2 Other 10.2 Professional and management consulting services 10.2.2 Advertising, market research, and public opinion polling 10.3 Technical, trade-related and other business services

Receipts

Payments

142 19 18

161 21 17

16

17

15

15

7

8

6

7

180

181

116 10 7 102 19 177 100 26

134 10 7 110 23 181 106 25

25

24

25

25

8 7

8 7

7

8

6 23 129 35

7 22 138 34

161

162

     



Table 5.3 (cont.) Receipts 10.3.1 Architectural, engineering, scientific and other technical services 10.3.1.1 Architectural services 10.3.1.2 Engineering services 10.3.1.3 Scientific and other technical services 11 Personal, cultural, and recreational services 11.1 Audiovisual and related services 11.1.1 Audiovisual services of which: 11.1.1. Audiovisual originals 11.1.2 Artistic related services 11.2 Other personal, cultural, and recreational services 11.2.1 Health services 11.2.2 Education services 11.2.3 Heritage and recreational services

Payments

35

35

25 26 24

24 25 24

142 89 7 6 7 116 26 25 23

154 96 7 6 7 121 26 26 24

Source: IMF.

services, now representing over 110 reporters for receipts and 130 for payments, followed by audiovisual (around 90). Surprisingly the number of reporters of advertising, market research and public opinion polling has markedly decreased, from 70 down to 35 reporters. However, it is important to note that for many of these items, developing countries, particularly those in Africa, do not report any data. While we have described above substantial improvements in the reporting of some main IP-related items, it is also important that we analyse further the level of development of statistics at the more detailed level introduced by MSITS2010. There, even though MSITS2010 and a number of other guidelines were released, and policymakers have referred to the need to obtain further details, the picture is rather disappointing. Only a very small number of reporters (mainly OECD or EU countries) have compiled information for those. As shown in Table 5.3, only 15–20 economies report details for charges for the use of intellectual property, 25 report details for research and development services (of which many are shown as having a zero value!), and fewer than 10 economies report data on sales or the transfer of proprietary rights. Finally, data broken down by counterparts in the transactions is available in 40–50 economies (mainly OECD and EU countries). Consequently, it



      

is important that users of the data and compilers further discuss the data gaps and identify ways to better respond to information needs. It is also important to state that IP-related trade can also take place through the temporary presence of natural persons (referred to as ‘mode 4’ delivery of services in the parlance of the General Agreement on Trade in Services). Very little information is available separately for mode 4, but the WTO Secretariat recently released an experimental dataset on Trade in Services by Modes of Supply (TISMOS).7 The economic transactions involving this mode of trade are indistinguishably covered in the balanceof-payments items indicated above. They are most likely important in the following items: computer services, professional and technical services (consultancy, advertising, architectural, artistic, education and health services (see Table 5.3)). When it comes to obtaining counts on mode 4 persons, the main area of interest will be on persons travelling abroad in the context of service contracts that they deliver either on their own behalf or on behalf of their employer. Again, very little data are available, but guidance is prepared on developing international labour mobility statistics.8 It is hoped that these guidelines will improve data availability enabling an analysis of mode 4, where in many cases an individual’s knowledge of the person is important in delivering the service. The distinction between the Extended Balance of Payments Services (EBOPS) 2010 categories from a reporter’s perspective may not always be clear (for example software versus research and development, or sales of IP rights versus licensing for use of IP). The suggested granularity of the breakdown of the R&D item is also often seen as challenging by compilers. When it comes to collecting the information, there are basically two main ways of gathering information: reporting by commercial banks of international transactions made by their clients (international transaction reporting system) and the use of surveys. The latter offers more options to collect detailed data as covered in this chapter. Given the level of detail sought and the complexity of the subject, one needs to clearly draft appropriate guidance for respondents, which may be particularly difficult when an international transaction reporting system is used (and probably also for general trade in services surveys). But the refined classification as presented in MSITS2010 should provide a solid point 7

8

An experimental data set was released on 31 July 2019; see ‘WTO launches trade in services dataset by sector and mode of supply’, available at www.wto.org/english/news_e/ news19_e/serv_31jul19_e.htm. See UNECE Guide on International Labour Mobility Statistics.

     



of departure for compilers. Interestingly, the US is spearheading developments on implementing these concepts. It has issued an enterprise survey collecting detailed information on international transactions relating to intellectual property.9 Also, recent developments in business strategies, and the blurring of activities is rendering the collection and compilation of information difficult. There are new major players involved (e.g. telecommunications companies, Internet service providers) in distributing IP products, which may contribute to the blurring of fields of activities. This may complicate the identification of the sample of enterprises to survey. There are also a few borderline issues concerning the rights acquired by transmitters such as TV companies and radio stations, when the contractual arrangements allow for multiple transmissions. Where the fees are paid on a ‘pay as you go’ basis, for example a royalty payment is made every time a song is played on a radio station, these should be recorded under audiovisual services while a one-off fee paid for unlimited air-time should be recorded as being equivalent to a licence to reproduce (charges for the use of intellectual property n.i.e.). Finally, although flows between enterprises (whether affiliated or unaffiliated) should be accounted for in financial statements, they may not always be identifiable as specific IPrelated transactions. As for the technological means of delivery, it is important to note that trade involving IP increasingly takes place online through e-commerce. The flow of IP-protected information is an item of increasing concern to service suppliers around the world, and, in particular, in relation to flows taking place through the Internet or other types of data networks. These data flows may reflect the actual service being provided (i.e. music being downloaded, film being streamed, blueprint), the transfer of a copy of the original on which the recipient will be entitled to make copies for distribution or sale, or refer to the information essential to service suppliers to be able to manage customer relations and data or to enable the international payments associated with sales, etc. Statistics on ITenabled transactions and e-commerce should also become an important complementary source of information when considering IP-related transactions.10 It is therefore important to take these developments into 9 10

For transaction types covered, see USBEA BE-120, ‘General Instructions’, p. 24. A related project on measuring trade in ideas attempts to categorize transactions related to IP to analyse further which of these transactions are available in data collection systems of national statistical offices to build statistical indicators.



      

account when developing strategies to improve the quality and relevance of statistics relating to IP international transactions.

Additional Measurement Challenges in GVCs The establishment of global value chains is a driving force for the increase of transactions in IP. International competition forces firms to increase their efficiency and develop new products. This phenomenon poses additional significant measurement challenges for the items identified in the previous section. Outside MNEs, IPP-related transfers or payments are tracked using market transactions and the identification of IPP ownership is usually less problematic. Yet, the analysis of IPP use in production still typically requires a complete picture of the global production chain (e.g. ‘factoryless goods producers’, ‘virtual manufacturing’). However, within MNEs it is more complicated. First, intra-firm transactions (i.e. within multinationals) may be affected by (i) the use of transfer pricing and/or more general tax planning issues which distorts the valuation of trade transactions, and (ii) decisions where to establish multinational affiliates or headquarters11. Transfer prices, being the cost at which assets are transacted between the divisions of one firm, may significantly differ from the actual ‘real’ market transaction values, negotiated by separate firms at arm’s-length from one another. Also, the distinction between flows recorded as trade in services and flows recorded as property income may not always be clear-cut. For example, R&D-related transactions or transactions relating to the use of an asset between affiliated enterprises are not always observable. They may also be channelled through convoluted chains of affiliates. Often, payments that are not explicitly recorded, but that are implicitly related to R&D or the use of an underlying asset are instead recorded as property income. In principle, an imputed payment should be recorded (either as a payment for R&D services or charges for the use of intellectual property n.i.e.). Hence, more often than not, it is unlikely that such imputations will be made in practice. Nevertheless, if compilers do make such imputations, they should ensure that counterpart transactions (in particular international transactions) should be coherent, requiring coordination with other statistics agencies. 11

See Thomas Neubig and Sacha Wunsch-Vincent, ‘A Missing Link in the Analysis of Global Value Chains: Cross-Border Flows of Intangible Assets, Taxation and Related Measurement Implications’, Chapter 6 in this volume.

     



Second, defining the location of the economic ownership of IP assets may be an additional barrier to the appropriate measurement of (international) transactions, with the fragmentation of activities, rendering transactions hardly identifiable.12 Indeed, multinationals may choose to register their patents, industrial processes or originals in one country rather than another based on ‘benefit maximizing’ considerations. Patents could be registered: • with the parent company; as the ultimate beneficiary, the parent company often financing directly or indirectly (via an affiliate) the acquisition or production of the IPP (i.e. the original) • with an affiliate using the IP in its production • with an affiliated intermediary which could be the original producer or could be a Special Purpose Entity (SPE) whose sole purpose is to register assets and act as the legal owner of IPPs, therefore obtaining the revenues of IPP copies or licences to use or reproduce. SPEs are usually established in view of optimizing the overall post-tax MNE profits. The country of registration will be where the legal ownership is established. But this does not mean that the country of registration is necessarily the same as that of the economic owner of the IP (i.e. the entity that draws benefits and may encounter losses from the ownership of the asset, whether the producer of the original or the entity providing or using the services or copies of the original). Not all countries may be able to clearly identify these aspects. As shown above, although recent international statistical standards clarified the classification of transactions, they failed to provide clear recommendations as to how to compile the information. When transactions are not observed directly, ownership and the recording of intragroup IPP transfers become uncertain and can only be based on certain 12

In national accounts and the balance-of-payments statistics, the recording of product transactions is done based on economic ownership change, which is a fundamental principle. Economic ownership is defined in para. 3.26 of the SNA as follows: ‘The economic owner of entities such as goods and services, natural resources, financial assets and liabilities is the institutional unit entitled to claim the benefits associated with the use of the entity in question in the course of an economic activity by virtue of accepting the associated risks’. As indicated in the Guide to Measuring Global Production, accepting associated risks involves the owner’s responsibility for maintenance of the asset, which can entail paying for fees to maintain patents, copyrights or other registrations of the IPP in question. The benefits can usually be derived from embedding IPP in its own production and therefore the subsequent sales of goods or services, or alternatively by licensing other parties for the right to use the IPP in their production processes.



      

conventions. The Guide on Measuring Global Production provides some guidelines in relation to global production arrangements, from the perspective of national accounts, as well from the perspective of the trade in services and balance-of-payments statistics. This complements the compilation guidance drafted by the UN expert group on the compilation of trade in services statistics. The Guide on Measuring Global Production provides a decision tree to determine the economic ownership of IPP and related IPP transactions. ‘The starting point is the observation of IPP output or IPP ownership at the level of a certain unit. The obtained information is examined in four different steps: 1. 2. 3. 4.

Control ownership of the unit: is the unit member of an MNE (yes/no)? Is the unit the producer of the IPP (yes/no)? What is the main kind of activity (in terms of ISIC) of the unit, or, is the unit expected to use the IPP in its production process (yes/no)? Does the unit receive income related to IPPs, or, does the unit pay for the use of IPPs (royalties and licences) (yes/no)?’

Of course, to be able to use such a decision tree, appropriate source data are needed. Nevertheless, compilers are strongly advised to implement this decision tree, as this should enable the compilation of relevant and internationally consistent data.13 Following the establishment of these more detailed guidelines, some specially targeted enterprise surveys should help improve the data situation when it comes to the more detailed information sought on IP, intra-firm transactions and location of ownership of IP (e.g. trade in service, FATS surveys, surveys used for the collection of R&D data). More research is needed on how to compile IPP-related transactions when SPEs are involved. A clear and separate identification of the IPP income flows received by IPP-holding SPEs should be classified under a separate heading such as ‘IPP related services provided by SPEs’ as the provision of these services does not really reflect economic transactions. Such a practice will help to support the correct interpretation of economic statistics where these entities are located, as well as help improve the data in other countries, whether the related MNEs are established there (parent companies or affiliates) or not, as well for countries where companies have unrelated/arm’s-length transactions. 13

Given the difficulties in compiling the information, it is not recommended to reroute the ownership of IP by SPEs.

     



However, many countries/compilers will not be able to collect such (detailed) information accurately. Also, the increase in the response burden should be acknowledged. A solution for these could be to complement their more general data collection systems with the information collected and disseminated by countries that will be engaging in a more detailed and sophisticated data collection system (most probably the countries which have interest in this information, that is the ones where research and development, and more broadly IP, is significant and where multinationals are involved). However, this can only function if there is efficient cooperation between compilers of different countries. In addition, it will be necessary to have detailed bilateral information published by countries that will be engaging in a detailed data collection system. Finally, an alternative global production arrangement is receiving increased attention in the statistical community. Although it is not a new channel, global producers and distributors are making rapid and increasing use of the opportunities offered by e-commerce. Consequently, the development of new business models for transactions in intangibles has been observed in recent years. This is described in international statistical guidelines as ‘merchanting of services’ (see MSITS2010 and UNECE Guide to Measuring Global Production). It can be described as follows: an intermediary located in one country arranges the supply of services/IP between suppliers and consumers located in other countries, without itself being engaged in operational activities. For example, a firm or establishment (i.e. the entity engaged in the merchanting of services) is created in a jurisdiction with the sole purpose of distributing copies of IP-protected material to other countries on behalf of a client in a third country. This is expected to be applied by MNES, but is not limited to those. At the time of writing, only an attempt has been made to conceptualize this practice for statistical purposes, but it seems that this way of doing business is of importance for several major service trading economies. Gathering further information on these types of transactions may also shed more light on how IP is traded.

Characteristics of IP Traders The first sections of this chapter deal with issues relating to the classification, collection and recording of information on international transactions related to IP. But how can we understand better the characteristics of those engaged in such transactions? How can we analyse the transactions with respect to the size of enterprises, their activities and the



      

relationships between supplying firms and recipients/clients? These are important questions for trade and other policymakers. They would need this type of information to improve their decision-making when it comes to determining export strategies as well as for trade negotiations. Trade by enterprise characteristics (TEC), compiled to complement the traditional merchandise and services trade statistics, can eventually respond to the questions raised in the previous paragraph. Data on international merchandise trade and trade in services are presented according to enterprise size class (measured in number of employees) and type of ownership (domestic whether with only local activities or multinationals and foreign-controlled affiliates). Data on merchandise trade by enterprise characteristics have been available for a number of years (collected in certain economies since 2006), and one would need to focus on high-technology product groups (e.g. aerospace, computers, electronics and telecommunications, pharmaceuticals, scientific instruments, chemicals), or relevant specific technology-intensive industries, to analyse merchandise trade related to IP. Statistics on Services Trade by Enterprise Characteristics (STEC) have received greater attention in recent years. Eurostat and the OECD recently released a compiler’s guide to produce such data.14 One important feature of such a dataset is that it does not require the addition of a response burden on reporters. However, one needs to have the necessary legal framework and data collection system in place enabling data linking between trade in services, micro data and enterprise characteristics. Keeping in mind confidentiality constraints, it would be useful to break down the information to a level of interest to analysts and policymakers. The level of detail recommended for the type of economic activity is aggregated and it will probably be difficult to suggest more detailed activities given the infancy of this new way of presenting trade statistics. Nevertheless, the current recommendations do suggest a level of disaggregation of services products that would be extremely useful in our context, as this would separately identify charges for the use of intellectual property n.i.e. and research and development services. Although these two items are of high relevance, other recommended items may also be useful (computer, audiovisual services, etc.). Finally, presenting the information by type of ownership would also be extremely useful to understand the role of multinationals in 14

See Compilers Guide for Statistics on Services Trade by Enterprise Characteristics (STEC), OECD and Eurostat, 2017 edition.

     



knowledge diffusion. This can either be analysed in the context of TEC and STEC (Services Trade by Enterprise Characteristics), as well as looking more closely at foreign direct investment statistics and FATS. For FDI statistics, it would be interesting to analyse foreign investment in knowledge-intensive activities, as well as to obtain a better understanding of the role of SPEs in that context. In the absence of statistics applying the economic ownership principle, it would be useful to gather more detailed information with respect to SPEs that are used to register assets and act as the legal owner of IPPs, therefore obtaining the revenues of IPP copies or licences to use or reproduce. Other sources that should be examined to obtain a better understanding of MNE operations with respect to IPPs are operational data portraying the activities of affiliates (sales, exports and imports, employment, assets) as captured by FATS. The information is presented according to the main sector of affiliates’ activities and enables the analysis of the performance of domestic- and foreign-controlled enterprises. It can be particularly interesting to obtain information pertaining to the eventual ownership of IP assets and transactions in IPPs, such as sales of products with embedded IP and licensing agreements, whether domestic or international (in the latter case, linking to balance-of-payments data).

Conclusion – A Way Forward This chapter has shown that trade in IP products has grown rapidly over the past decades and at an accelerating pace in recent years, driven by the development of global production arrangements, as well as the digitization of the economy. While concepts in statistical frameworks have been clarified, data collection and compilation do not keep pace. At the time of writing we are far from having a full picture of international transactions pertaining to IP. Main BPM6/MSITS2010 items are compiled (computer services, audiovisual services, charges for the use of intellectual property, n.i.e.) but, in most cases, we do not have the necessary details (let alone breakdowns by trading partner) to conduct a thorough analysis of international IP transactions. The only item that presents a notable recent increase in reporting is research and development services. Also, as the section entitled ‘Additional Measurement Challenges in GVCS’ has shown, several challenges remain when it comes to compiling these items. As has been outlined, one of the main issues is the valuation of intra-firm transactions, which is particularly striking for IP-related transactions. The other important issue pertains



      

to the identification of the ownership of IP rights, and more particularly to the economic relevance of SPEs being identified as owners of IP rights in economic statistics. Finally, the section entitled ‘Characteristics of IP Traders’ has highlighted the need to obtain more information for better trade and IP policy analysis, and decision-making (size of firms, relationship between firms, type of global production arrangements etc.). Based on the points raised in this chapter, we suggest some steps that could be explored to improve the data situation: • It is advised to implement BPM6 recommendations, including for the services items of interest by all reporting jurisdictions. Only this would allow implementation of the level of detail recommended in MSITS2010 (including data broken down by partner). • Find synergies between relevant data collection and compilation methods based on existing compilation guidelines (BPM6 Compilation Guide, MSITS2010 Compilers Guide, UNECE Guide on Measuring Global Production, etc.) with mechanisms established for other related statistical needs (audiovisual, culture, creativity, e-commerce). Specific questions should be added in (trade in services) surveys, and if deemed necessary, specific surveys could be developed. Innovative sources of information could also be explored (e.g. big data). • It is crucial to better understand the (emerging) business models to address compilation issues (ownership of IP rights, who trades with whom, intermediary platforms in e-commerce, etc.). Existing international guidance, such as the Frascati Manual and the UNECE Guide to Measuring Global Production, should be followed. More information on SPEs and their role in global production arrangements needs to be gathered (including through FDI statistics and FATS). • Given the importance of knowing more about the characteristics of the firms that trade in IP products, it is suggested that ways of developing statistics on trade broken down according to enterprise characteristics, focusing on items of interest in the context of IP (computer, audiovisual services etc.) should be found. • This cannot be done efficiently without the promotion of cooperation at the national level for defining priorities and needs, and finding synergies for data collection and compilation (i.e. need for coherence and reducing the burden of data collection). This needs to be accompanied by a policy push if it is to succeed (engaging those authorities

     



and policymakers that deal with trade, IP, innovation and related industry and technology policies). There should also be cooperation at the international level to ensure the comparability of the data, to promote and assist in the development of relevant statistics, and to help compilers exchange information and experiences, such as data-sharing arrangements and further guidelines based on practical experience. To respond to information needs, there is also the need to create a new international dataset covering international IP-related flows (including relevant services items as well as merchandise statistics to cover embedded IP). For example, work could be done in the context of the OECD– WTO Trade in Value Added (TiVA) initiative to better measure IP embedded in traded products.

Bibliography OECD and Eurostat, Compiler’s Guide for Statistics on Services Trade by Enterprise Characteristics (STEC), 2017 edition. UNECE, Guide on International Labour Mobility Statistics, Geneva and New York, 2018. WTO, World Trade Statistical Review, Geneva: WTO, 2018.

6 A Missing Link in the Analysis of Global Value Chains: Cross-Border Flows of Intangible Assets, Taxation and Related Measurement Implications                          -        * Abstract Global fragmented production and associated trade in intermediate products has changed how economists study globalization and how new public policies are shaped. This creates a need for understanding cross-border flows of disembodied knowledge, often associated with intellectual property (IP). However, data on these international IP-related knowledge flows, specifically, cross-border payments for use of IP, are distorted. The chapter shows how tax planning by multinational enterprises has seriously distorted the measurement of cross-border IP flows, with tax-induced mismeasurement potentially exceeding 35% of global Charges for Use of Intellectual Property (CUIP), and more for individual countries, particularly high-taxrate countries. This chapter describes the trend of increased cross-border trade in ideas, assessing the main data sources and distortions to these flows. The chapter analyses the tax effects on cross border IP measures and the magnitudes and trends in the tax distortions. It concludes with some possible avenues for reducing mismeasurement distortions, directed at national statisticians, national tax administrations, policymakers and analysts, and academics.

Introduction The phenomenon of global fragmented production and associated trade in intermediate products, including intangible assets, has changed how economists study globalization and how new public policies are shaped. In this light, understanding cross-border flows of disembodied knowledge, often associated with intellectual property (IP), is essential to analysing how modern economies operate. Unfortunately, available data * The views expressed in this article are those of the authors and do not necessarily reflect the views of the World Intellectual Property Organization or its member states.



               



to document these international IP-related knowledge flows – namely cross-border payments for IP1 – are distorted. Indeed, IP-related tax avoidance strategies have biased measures of trade, GDP and productivity. As this chapter shows, tax planning by multinational enterprises has seriously distorted the measurement of cross-border IP flows and affected the national measurement of imports, exports, GDP and productivity. The tax-induced mismeasurement could be more than 35 per cent of global Charges for Use of Intellectual Property (CUIP) and more for individual countries, particularly high-tax-rate countries. The chapter is structured as follows. In the first section, the trend of increased cross-border trade in ideas is described. The main data sources assessing these flows are portrayed in the second section. In the third section, the main distortions to these flows are assessed. Then, in more detail, in the fourth section, the tax effects on cross-border IP measures are described and the magnitudes and trends in the tax distortions are presented. Finally, in the last section, the chapter discusses some initial possible approaches to reducing mismeasurement distortions.2

The Rise of Cross-Border Trade in IP Technology, business innovations, and falling trade costs have transformed the organization of global production. The unbundling of the production process and the geographical dispersion of different stages of production are the key elements of this transformation. Increasingly, multinational enterprises (MNEs) source input and technology from suppliers worldwide. This reflects a fragmentation of the production process in the manufacturing and services industries, with increases in task-based manufacturing, intermediate trade and the outsourcing of services. As a result, a greater number of countries participate in global production and innovation networks.3 Manufacturers exposed to these networks experience technological and organizational learning, possibly leading to industrial upgrading. 1

2

3

See Joscelyn Magdeleine and Andreas Maurer, ‘Measuring International Intellectual Property Transactions in a Globalized World: Current Challenges and Possible Improvements’, Chapter 5 in this volume. The underlying findings of this chapter were produced to inform the WIPO World Intellectual Property Report 2017, on ‘Intangible Capital in Global Value Chains’, WIPO (2017). WIPO (2011).



    - 

It is now a common assumption that intangible assets in the form of technological knowledge, software and other types of know-how are key in the analysis of increasingly fragmented production. The causes and consequences of this cross-border trade in ideas are manifold: • As firms focus on their core competitive advantage, they purchase essential technologies from third parties rather than developing themselves, also propelling the vertical disintegration of knowledge-based industries.4 • The increasing existence of ‘technology markets’, backed by (i) new business models such as IP licensing, (ii) information technologies, (iii) more legal and practical experience with technology transactions, and (iv) increased IP protection worldwide allow for such technology transactions to increasingly take place. • The operations of MNEs have led to more intra-company transfers of technology, also because of increased foreign direct investment. • Factoryless production is on the rise in which firms outsource their manufacturing activities but control the underlying IP (including the brand name) and hence control the production and value chain outcome. • The mix of productive assets in business has shifted decisively since 1990 to the point where a majority of capital investment in developed economies is in technology, software, design, creative works, brands, human capital and business organization. About half of these intangible investments are ‘owned’ via formal IP rights.5 These factors have spurred an increase in both inter-company and intra-company trade in disembodied know-how. Such knowledge is frequently subject to registered IP, such as patents and industrial designs, and unregistered IP, such as copyright and trade secrets. As a result, ensuring the accurate measurement of exports and imports of the trade in IP is essential.6 Importantly, the types of market-based transactions of IP include the outright sale of originals, the temporary licensing of copies/technology,

4 5

6

See Arora et al. (2001) and WIPO (2011), chapter 2. Estimates of business intangible investment for the United Kingdom and the role of IP rights are estimates in work commissioned by the UK IPO, see https://assets.publishing.service .gov.uk/government/uploads/system/uploads/attachment_data/file/554480/Investment-inIntangibles.pdf. OECD (2014).

               



but also more complex arrangements such as cross-licensing deals, patent pools and joint R&D ventures or cost-sharing arrangements. For measurement purposes these exchanges can take various formalized, market-based transactions of which some entail explicit pecuniary transactions – including notional payments for intra-corporate IP transfers among headquarters and company affiliates – and others involve the sharing of technology without explicit payment from buyer to the seller (see Table 6.1 for a taxonomy). In all likelihood, some of the below transaction types involve self-declared estimates or are unrecorded in Table 6.1 Taxonomy of formalized and informal technology transactions Financial transaction status and degree of documentation Formalized IP-based transactions or sharing agreements

Informal technology sharing

Pecuniary and recorded

Nonpecuniary but recorded Nonpecuniary and unrecorded

Type of transfer Market-based transactions with one buyer paying seller. Transactions among headquarters and affiliates involving notional estimates of the transaction. Mutual sharing of technology. Non-market-based use of third-party technology with no associated payment. IP infringement.

Legal status of transfer Commercial licensing of a specific technology with contract and payment/sale or purchase of a patented industrial process. Intra-company transfers. Cross-licensing, patent pools. Implicit, unwritten cross-licensing arrangements or covenants not to sue. Consent decrees. Unrecorded intracompany transfers. Patent infringement or piracy.



    - 

existing statistics, hence biasing estimates of trade in IP, mostly downward. It is important to recognize that only (i) pecuniary, market-based, authorized transactions or (ii) self-declared, notionally valued IP transactions are captured by official statistics. Even in these two cases, the proper valuation of licensing transactions is complex, leaving room for arbitrariness or misreporting reflecting tax or other considerations.

Available Data on Cross-Border IP-Related Flows Two official sources exist to trace cross-border IP flows, namely (i) international trade data that set out cross-border receipts and payments for the purchase or use of IP, and (ii) tax data. As regards the former, these international trades in services data are now commonly available through the balance-of-payment (BoP) statistics at the level of countries collected by the International Monetary Fund (IMF) and available in the World Development Indicators (WDI) database, and at the level of the OECD – and a few non-OECD economies – as technology balance-of-payment data. In addition, a limited number of specific countries publish these data in a more fine-grained fashion – also capturing intra-company transfers – based on specific quarterly, annual or other surveys. The United States (US), in particular, has been publishing detailed intra- and inter-firm data on IP flows for some time. Only a few national data sources allow for such a breakdown, namely Canada, Finland, Israel, Italy, Japan, the Netherlands, Poland, Sweden and the US. Yet, even for these few highincome economies, not all of these data are detailed enough for the analysis of intra-corporate IP flows. The most widely reported metric on disembodied technology trade relates to the international receipts and payments for the ‘authorized use of intangible assets and proprietary rights’.7 One advantage of these data is that most countries publish them in a timely and regular (yearly or quarterly) manner. In recent years, significant progress has been achieved 7

The IMF defines royalties licences and fees (RLF) as including ‘international payments and receipts for the authorized use of intangible, non-produced, non-financial assets and proprietary rights . . . and with the use, through licensing agreements, of produced originals or prototypes . . .’. See International Monetary Fund, Balance of Payments Statistics Yearbook and data files, available at https://databank.worldbank.org/metadata glossary/africa-development-indicators/series/BX.GSR.ROYL.CD.

               



2.0

450

1.8

400

1.6

350

1.4

300

1.2 250 1.0 200 0.8 150

0.6

100

0.4

50

0.0

0

1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

0.2

IP payments (BoP, current US$ bn)

IP receipts (BOP, current US$ bn)

IP payments as % share of imports

IP receipts as % share of exports

Figure 6.1 Royalties Licences and Fees (RLF) receipts and payments, in USD billion and as share of trade Source: Authors based on World Development Indicators.

to deliver more reliable data on cross-border technology payments (IMF, 2009, UN et al., 2011). Steps have been taken to improve international accounting standards of national accounts and BoP to better record transactions of firms participating in global production (de Haan et al., 2014, and UN ECE, 2015). The associated data show that the rise of global value chains (GVC) has coincided with increasing cross-border trade in ideas and IP across countries since the 1990s in high-income countries and then as of the 2000s in upper-middle-income countries mostly through Asian economies, particularly China.8 In 1990, 62 countries reported making licensing payments but by 2015, this number had increased to more than 120 countries (see Figure 6.1). The US has the largest amount of receipts (exports) followed by the Netherlands and Japan. Ireland has the largest amount of payments 8

See Athreye et al. (2012), WIPO (2011a) and IMF (2011).



    - 

(imports) followed by the Netherlands and the US. The US and Japan are the largest net exporters of IP services while Ireland and China are the largest net importers. A number of tax havens do not report RLF receipts, but there are other reasons for this discrepancy discussed below. The OECD’s Technology Balance of Payments (TBP) follows the same methods as for the IMF BoP manuals.9 The magnitude of Ireland’s surplus in technology receipts stands out. It is known that the strong presence of foreign affiliates (particularly US and UK firms) and related intra-firm transactions play a role in Ireland, potentially also affected by tax transfer pricing of MNE internal-related party transactions. For countries where detailed data are available, it is known that IPrelated payments mostly consist of intra-firm payments, i.e. payments between subsidiaries and company headquarters. The US Bureau of Economic Analysis (BEA) data show that the majority of IP payments and thus imports – slightly more than 70 per cent of all recorded payments in 2015 – consist of intra-firm payments. Another government source of royalty data are derived from corporate tax returns. For instance, the US Internal Revenue Service (IRS) Statistics of Income publishes annual corporate tax statistics, which include royalties received by industry. The Australian Taxation Office requires certain taxpayers to complete an annual international dealings schedule, which contains information on international related party transactions, which can include dealings in intangible assets. That said, corporate tax data are not as comprehensive as data from administrative surveys, since some royalties will be received or paid by non-corporations.

Key Distortions The statistical measures of cross-border IP flows are distorted by a number of factors.10 This section identifies some of the key distortions.

Incomplete Reporting Global flows of IP, measured in the Balance of Payments as Charges for the use of intellectual property (CUIP), not included elsewhere, totalled 9 10

OECD (1990). Some of the most important distortions were initially reported in Madeuf (1984) and OECD (1995).

               



USD 319 billion receipts and USD 359 billion payments in 2015. The USD 40 billion discrepancy between total receipts and payments indicates a statistical problem. Many countries do not report CUIP receipts and/or payments, thus contributing to an undercount. Countries such as the Bahamas, Bermuda and the Cayman Islands, which are home to many MNE affiliates with IP-holding structures, have not reported royalty receipts. Some of the other under-reporting distortions include: • Incomplete surveying: frequently, only firms with R&D activities are surveyed for their royalty receipts and payments, whereas other companies might also license in technology, hence leading to an underestimation (OECD, 2005).11 • Difficulty of separating disembodied from embodied technology flows: as described in OECD (2005), intangible IP flows often take place in conjunction with, or in addition to, high-technology exports or imports or direct investments. While they ought to be reported as technology payments or receipts, in principal, in reality they are often difficult to identify separately from a firm’s other transactions. Payment for intangibles via channels other than technology payments: • if payment is made through channels other than technology payments, e.g. payments in the form of profits, dividends or overcharging for capital goods, then the payments for intangibles are not recorded appropriately (OECD, 2005). • Disaggregation: international transfers of IP ‘bundles’ which generate economic value can often be assessed as individual rights which on their own have lesser total value. The individual rights can be reassembled at their destination. Complementarities between individual IP rights can significantly increase the value of the individual holdings. • Tacit and non-pecuniary IP flows, and, of course, straight-out infringements of IP rights of third parties, which are unauthorized, result in undercounting of cross-border IP flows.

11

In some countries, surveys concerning the TBP are combined with R&D surveys (same sample). So the collected data could be underestimated, especially as far as payments are concerned, since firms without any R&D of their own which import technology from abroad are not taken into account.



    - 

Non-Valuation Measurement Issues Individual countries’ national statistics could reflect IP activity that could exaggerate the amount of real economic activity occurring in some countries, while understating it in other countries. Four types of issues are described below. • First, Special Purpose Entities (SPE) are often used for IP or financial holding companies which can increase measured inflows and outflows, as a result of serving as an intermediary or flow-through.12 • Second, characterization of cross-border IP flows can affect measured CUIP. While imports and exports of CUIP should net, if there is transformation in the character of the payments, such as from royalties to dividends, on one side of the transaction, then the relative IP position of countries would be distorted. • Third, profits from IP associated with contract manufacturing undertaken in other countries can increase the exports from the country of the contractor. • Fourth, combining the provision of services (annual flows) with the acquisition of capital investments (stocks) can provide a misleading picture of a country’s IP capabilities.

Valuation and Tax Issues One of the important IP measurement issues results from multinational enterprises’ tax planning and organizational structures. IP often plays a critical role in this planning. Favourable government policies, including lower tax rates for certain tax IP income, so-called ‘patent boxes’, increase the likelihood of tax-induced investment and potential profit-shifting. The amount and valuation of cross-border IP flows is affected by countries’ tax rules. Illegally evading taxes often results in undercounting of the related economic activity. Differential tax treatment of royalties and dividend payments may result in changing the character of payments to reduce taxes. More important are the distortions from the mispricing of transactions between related parties in different taxing jurisdictions to reduce a MNE’s global tax liability. Lower effective tax rates on IP investments can attract additional real investment in research and development. Low tax rates can encourage 12

Rassier (2015).

               



more registrations of patents in countries. Low marginal tax rates can encourage multinational enterprises to shift taxable income and reported financial profit from higher-tax-rate countries to lower-tax-rate countries. These tax effects, particularly base erosion and profit-shifting effects, can distort the measurement of cross-border intellectual property receipts, payments and assets. Several specific tax factors can distort national statistics from IP. The most well-known, and possibly the most important, factor is transfer mispricing. Transfer mispricing is where non-arm’s-length valuations are used for related party transactions. Arm’s-length prices between unrelated entities are often not available as comparables for valuation of related party transactions due to the unique nature of many intangibles. This can distort national statistics towards higher valuations in low-taxrate countries and lower valuations in high-tax-rate countries.13 The presence of tax-motivated transfer mispricing is found in numerous empirical studies, and is particularly strong among R&D-intensive firms.14 Beer et al. (2018) report the rate of the percentage change in the transfer price in response to a one-percentage point change in a country’s tax rate ranges from 0.5 to 6.0, with intangible transactions at the higher end. Another tax planning approach is the strategic location of valuable IP in low-tax-rate affiliates. MNEs can conduct their research and development activities in one country, but transfer the ownership of the patent or other IP to a country where the future income streams will be taxed at a low tax rate. Since valuation of transfers of unique IP is difficult due to lack of comparable transactions, especially when the IP is not fully developed, significant future profits can be shifted to low-tax-rate jurisdictions. Several studies report high responsiveness of patent and trademark applications to tax rate differences.15 Karkinsy and Reidel (2012) find a one-percentage point increase in the corporate tax rate reduces the probability of patent applications in a country by 3.5 per cent. The responsiveness of the location of intangible income to tax rates is greater when the location of the future income can be separated from the location of the actual value creation or R&D activity. A number of 13 14

15

Guvenen et al. (2017). Clausing (2003), Christea and Nguyen (2016), Hebous and Johannesen (2015), Liu et al. (2017). Grubert (2003), Dischinger and Riedel (2011), Karkinsky and Riedel (2012), Griffith et al. (2014), Dudar and Voget (2016).



    - 

countries have introduced so-called ‘patent box’ tax rules providing lower tax rates for certain types of intangible income, compared to their ordinary corporate tax rates. Studies found that patent box tax rates have a greater location effect on reporting of intangible income, but much less effect on actual local R&D activity, when the tax rules don’t require an ‘economic nexus’ linking actual R&D activity with intangible income.16 The OECD/G20 Base Erosion and Profit-shifting (BEPS) project, and the subsequent BEPS Inclusive Framework, requires the 115 member countries to have an economic nexus requirement on patent box tax regimes to prevent harmful tax competition.17 Several other potential distortions can result from the quantity of activity being mismeasured, even if the prices used are market prices. Structuring of business operations, entities and transactions may result in less measured economic activities in certain countries. In some cases, the structuring may facilitate transfer mispricing, but may result from tax minimizing strategies to avoid income tax in a jurisdiction, minimize withholding taxes or simply take advantage of lower tax rates.

The Location of Cross-Border IP Receipts and Tax Rates The OECD/G20 BEPS project addressed the disconnect between the location of taxable income and the location of value creation. The project instituted standards to prevent harmful tax competition, updated standards for tax treaties and transfer pricing, and instituted new country-bycountry reporting of MNEs for improved transfer pricing administration, among other changes. Information from the World Development Indicators on cross-border Charges for the Use of Intellectual Property (CUIP) has not previously been used to assess profit-shifting from transfer mispricing and the strategic location of IP, which have been estimated to be the main channels of base erosion and profit-shifting. Cross-border IP receipts should be strongly related to prior R&D expenditures. A simple analysis examines whether corporate tax rates, including tax rates for certain intangible income, affect the relationship of cross-border IP receipts and prior year R&D expenditures across countries. Figure 6.2 shows a high ratio for several countries with low tax rates or tax regimes 16 17

Evers et al. (2013) and European Commission (2015). See www.oecd.org/tax/beps/beps-about.htm.

               



300% 250% 200%

150% 100% 50% 0%

Figure 6.2 Charges for the use of IP as a percentage of lagged R&D, 2016 Source: World Bank World Development Indicators (datacatalog.worldbank.org/dataset/worlddevelopment-indicators) and authors’ calculations.

providing low effective taxation of IP. Luxembourg, Netherlands and Ireland have ratios exceeding 175 per cent while Japan, Germany and the US have ratios below 25 per cent. Regression analysis of CUIP, R&D, outbound FDI and tax rate data for 29 OECD countries from 2005 to 2016 finds that a one-percentage point lower tax rate on IP income increases the ratio of IP income to lagged R&D expenditures by 5–7 per cent (Table 6.2). This is consistent with other empirical findings showing greater transfer mispricing in R&D-intensive sectors and the strategic location of IP in countries with low tax rates. It is unlikely that non-tax explanations can fully account for the consistent pattern of higher IP receipts relative to prior year R&D expenditures recorded in countries with lower tax rates. Future research could use this type of data to determine if the OECD/G20 BEPS project and implementation of anti-avoidance rules by countries since 2014 are reducing tax-motivated distortions in the IP flows and national statistics. The regression finds the general corporate tax rate to be statistically significant, but there is a larger effect when the intangible tax rate is used. The differential between the general and intangible tax rate has a larger effect than the intangible rate by itself. The ratio of countries’ outbound foreign direct investment (FDI) to GDP is positive and statistically significant. High relative outbound FDI is expected to be associated with



    - 

Table 6.2 Regression results of excess cross-border receipts to tax rates Dependent variables CUIP/ R&D

CUIP/ R&D

CUIP/ R&D

In (CUIP/ R&D)

In (CUIP/ R&D)

In (CUIP/ GDP)

Intercept

0.517 0.092

0.652 0.063

0.588 0.087

1.142 0.249

1.669 0.344

5.963 0.299

General corporate tax rate

0.012

0.016

0.031

0.066

0.003

0.003

0.013

0.011

0.021

0.065

0.067

0.003

0.012

0.01

Explanatory variables

Intangible tax rate

0.019

0.049

0.002

0.009

Differential between general & intangible tax Outbound FDI % of GDP

0.018

0.014

0.013

0.028

0.025

0.024

0.001

0.001

0.001

0.006

0.006

0.005

In (R&Dt-1/ GDPt-1)

2.275 0.118

R2 adj. N

0.353 348

0.429 348

0.429 348

0.199 348

0.208 348

0.585 348

Source: World Bank World Development Indicators (datacatalog.worldbank.org/ dataset/world-development-indicators) and authors’ calculations.

               



greater cross-border activity, but also reflects countries with significant SPEs, which often facilitate tax planning. A number of alternative specifications were tested. The last column shows when CUIP as a percentage of GDP is an absolute rather than a ratio to lagged R&D expenditures, with a slightly larger tax differential effect. Time fixed effects don’t affect the size of the coefficients. The coefficient on the tax rate differential variable declines with country fixed effects, suggesting most of the effect occurs across countries rather than within individual countries.

Magnitudes and Trends in Tax Distortions of Cross-Border Flows in National Statistics Ideally there would be a comprehensive database of unconsolidated firmlevel data over multiple years with which to econometrically estimate the effects of corporate tax rate differentials on profit-shifting. Such a database doesn’t exist, so researchers have turned to aggregate national statistics as well as available firm-level databases. The BEPS project requires large MNEs to file country-by-country reports (CbCR) of their economic activity and taxes by individual firms with tax administrations. The OECD will provide aggregated tables of the CbCRs starting in 2019. The most-used global firm-level database, ORBIS, has millions of financial account records for consolidated and unconsolidated entities globally, both domestic only and multinational enterprises. However, the database is missing companies from many parts of the world, since MNEs headquartered in Europe accounted for 69 per cent of the affiliated firms in the database, and of the affiliates with key financial information 78 per cent were located in Europe. Thus, the coverage of the US, BRIC countries and low-income countries is far from complete. Further, some of the affiliates that had been flagged in parliamentary inquiries for large amounts of profit-shifting were not included in the database or did not include financial information.18 Further, to estimate the revenue consequences of profit-shifting, only financial data, not tax return, information was available. Researchers have turned to several types of country-level aggregate data sources to assess the general magnitude of profit‑shifting, or have used elasticity estimates from individual firm data matched with countrylevel aggregate data. Table 6.3 shows seven estimates of the annual global 18

OECD BEPS Action 11 report (2015).



    - 

Table 6.3 Estimates of global and low-income country fiscal effects from profit-shifting Magnitude (billions)

Fiscal estimate approach

Scope

OECD Aggregate tax rate differential 2015

Global

100–240 (4–10% of CIT)

IMF CIT efficiency 2014 UNCTAD Offshore investment matrix 2015

Global Global

5% of CIT 200 (8% of CIT)

Clausing excess income in low tax countries 2016

Global

Cobham and Jansky tax haven spillover 2018

Global

Crivelli, Keen and DeMooij tax haven spillover 2016 Torslov, Weir and Zucman high profit-to-wage ratios of foreign-owned firms 2018

Global

279 (20% of CIT) 494 (24% of CIT) 647 (32% of CIT) 10% of CIT

IMF CIT efficiency 2014 UNCTAD offshore investment matrix 2015

Global Developing countries Developing countries

13% of CIT 66–120 (7.5–14% of CIT)

revenue loss, and two estimates of annual losses in low-income countries, from base erosion and profit-shifting. These estimates were done at an aggregated basis so do not separate profit-shifting from transfer mispricing and the strategic location of intangible assets from the strategic location of debt and other profitshifting tax minimization strategies,19 although transfer pricing and the strategic location of intangibles have been estimated to be the primary drivers of BEPS. What is clear is that profit-shifting is occurring as a result of multiple tax minimization strategies of MNEs and does affect the measurement of cross-border IP flows and trade in services, both in quantity and valuation. 19

The BEPS Action 11 report (2015) provides detailed information about these different analyses and the limitations of such estimates.

               



Estimates at the global or regional basis have to rely on aggregate country statistics, since individual firm information across countries is incomplete. More detailed analysis using administrative records, including tax return information, could potentially provide estimates with a smaller range and greater reliability, but would still be estimates, since profit‑shifting has to be separated from shifts in real economic activity. To provide a general magnitude of the potential mismeasurement of cross-border IP flows for this chapter, we suggest starting from a global net revenue loss of overall profit-shifting. Table 6.3 shows a wide range from 4 to 32 per cent of global corporate tax revenues lost due to BEPS. Starting from a 10 per cent estimate of roughly USD 250 billion in 2015, the total loss can be shared down to the proportion attributable to intellectual property (see Table 6.4). The revenue loss from IP is then grossed back up to the magnitude of CUIP affected. Based on a number of critical assumptions, the amount of profit-shifting from cross-border CUIP flows in 2015 could conservatively be in the order of USD 120 billion annually. This results in an understatement by at least 35 per cent of the total trade in CUIP. This profit-shifting represents a disconnect between the location of measured IP profits and the location of where the economic activity generating the IP resides. The starting point of total global net revenue lost from base erosion and profit-shifting of USD 250 billion (line 1 in Table 6.4) assumes that Table 6.4 Estimate of global profit-shifting of cross-border CUIP flows USD billions 250 -83

1. Net global revenue loss from total profit-shifting, 2015 2. Exclude losses from tax shifting of interest income/expense (33%) 3. Net global revenue loss from non-interest profit-shifting 4. Share of net global revenue loss from CUIP before adjustments (3%) 5. Adjustment for higher profit-shifting elasticities for IP (2x)

+5

6. Net global revenue loss from CUIP profit-shifting 7. Average tax rate differential, weighted by trade flows

10 8.3%

8. Global net profit-shifting of cross-border CUIP flows 9. Average of exports and imports in CUIP, 2015

121 349

10. Net profit-shifting as per cent of trade in CUIP

35%

Source: Calculations by authors, described in text.

168 5



    - 

profit-shifting has continued to increase since the levels earlier in the decade on which the empirical estimates were based. The estimate assumes (line 2) that roughly one-third of the total global revenue loss is from the strategic location of both internal and external debt, consistent with some firm-specific analyses. This reduces the potential revenue loss from non-interest profit-shifting to USD 168 billion (line 3). CUIP is only a small share of total international trade. Although the distinction between goods and services is increasingly blurred, roughly 70 per cent of international trade is in goods, not services. Data on the underlying detail of trade in services in the EU28 countries and the United States shows that CUIP, not included elsewhere in national statistics, was only eight per cent of total trade in services. This leaves CUIP as only 3 per cent of total measured trade subject to potential transfer mispricing and other non-interest profit-shifting strategies. Assuming that CUIP has the same potential for profit-shifting as non-CUIP trade, then there would a global annual net revenue loss of USD 5 billion associated with CUIP (line 4), but that is an unreasonable assumption. Empirical research has found profit-shifting to be significantly higher in intangibles-related activity, as described above. The responsiveness of cross-border IP flows to profit-shifting from corporate tax rate differentials is assumed to be two times the responsiveness of other profitshifting, an elasticity of at least -2.0. This adjustment increases the annual global revenue loss from cross-border CUIP flows to USD 10 billion (lines 5 and 6). Any estimate of the potential government revenue lost from profitshifting depends on the tax rate differential between the countries from which and to which the profit was shifted. The amount of shifted profit will thus be the annual revenue loss from profit-shifting related to IP divided by the appropriate weighted average tax rate differential. Assuming the tax rate differential on IP flows is 8.3 percentage points20 (line 7) results in an estimate at around USD 120 billion CUIP flows annually (line 8) or about 35 per cent of total trade in CUIP, at 2015 levels (lines 9 and 10). Modifications to any of these assumptions would change the estimate, either higher or lower, but we believe that this is a conservative estimate of the understatement of global CUIP. It should be noted that while profit-shifting is zero-sum globally, the mismeasurement of global CUIP is unlikely to be zero-sum, and highly

20

BEPS Action 11 report (2015), p. 206.

               



likely to be understated. Fetzer et al. (2017) note that CUIP is a major area of asymmetry in trade in services statistics from national statistical agencies,21 which is apparent in the large discrepancy between total global receipts and payments of CUIP in Table 6.3. Tax minimization by MNEs in related party transactions results in overcharges for imported services to high-tax-rate countries (payments) and undercharges for exported services (receipts), and the reverse for low-tax-rate countries.

Mismeasurement for Individual Countries Given the current data limitations of measuring tax-induced profitshifting,22 attempting to estimate the CUIP mismeasurement for individual countries would not be reliable. However, the direction of the biases is clear, with understatement of net exports by high-tax-rate countries and overstatement of net imports by low-tax-rate countries. Several of the recent estimates of global revenue loss from profit-shifting have also made estimates of overall profit-shifting for individual countries, but the mix of IP and non-IP tax distortions is likely to vary by country. Using US BEA annual surveys of US-based multinational firms and their affiliated firms abroad, Clausing (2016) estimated – econometrically – profit-shifting from the US to other countries. She found that 82 per cent of the ‘excess’ income is booked in seven low-tax-rate (‘tax haven’) countries: Bermuda, Cayman Islands, Ireland, Luxembourg, the Netherlands, Singapore and Switzerland. Extending her analysis to the global profits of Global 2000 firms, she estimated that higher-tax-rate headquarter countries booked USD1.1 trillion of profits in low-tax-rate countries in 2012. Dowd et al. (2017) report significant profit-shifting into tax haven countries by US multinationals based on tax return data. They estimate that profit-shifting between foreign affiliates is highly non-linear with respect to corporate tax rates with very high elasticities in shifting to tax havens. Guvenen et al. (2017) estimate, based on US MNE data and an assumption that profit is related to sales, that profit-shifting reduces earnings on direct investment abroad of US multinationals by 65 per cent or USD 280 billion, with most coming from tax havens and from R&D-intensive industries. 21 22

Fetzer et al. (2017), p. 13. OECD BEPS Action 11 report (2015).



    - 

Main Trends in Cross-Border IP Profit-Shifting Estimates of the tax losses from profit-shifting have not been made across time; although there are some indications that profit-shifting has been increasing over time. Grubert (2003) examined tax return and other data for US MNEs and found higher levels of profit-shifting over time. Guvenen et al. (2017) also find a significant increase in profit-shifting by US MNEs since 2000. Clausing (2016) estimated the trend in revenue loss due to income shifting from the United States, showing a rapid rise after 2001 from under USD 20 billion to over USD 110 billion in 2012. Her estimated trend assumes a constant elasticity of profit-shifting to corporate tax differentials, so the trend reflects the doubling of income of foreign affiliates between 2004 and 2012 plus a continuing decline in average foreign effective tax rates, which are the incentive to shift profits out of the US. The variation in corporate tax rates among OECD countries has increased significantly between 2000 and 2017. An important question is whether significant changes in the international tax environment and national governments’ tax policies and administration will reverse what appears to be the prior trend toward more profit-shifting, particularly affecting cross-border IP flows. Tax authorities are focusing on these issues much more closely, as are more academic researchers, which will assist national account statisticians in improving their measurements.

Possible Approaches to Improve Measures of Cross-Border IP Flows Recent global and national initiatives in the areas of tax policy and tax administration are expected to reduce the amount of tax-induced profitshifting, and thus will reduce future distortions in the measures of crossborder IP flows. Additional steps by various stakeholders could further improve national statistics of cross-border IP flows.

Potential Actions of National Statisticians 1. 2.

ensure implementation of improvements in data on better classification and new IP flow disaggregation further extension of the work on SPEs as part of the OECD’s Benchmark Definition of Foreign Direct Investment and IMF’s BPM6, plus improvements in the measurement of intangible investments, including capitalization of investments

               

3.

4.

5.

6.



closer cooperation with tax authorities, and qualified academic researchers, on economic activities of MNEs, including potential agreements to secure better information from tax return data, as is currently done in some countries where possible, linking tax return data at the firm level with business financial account data would provide important insights. For example, linked data could be used to better evaluate the effects of tax patent box claims on the effectiveness on R&D and revenue impacts more focus on MNE activity, both inbound and outbound, including consideration of cost-beneficial imputations in satellite accounts. Additional work on identifying appropriate imputation methodologies is needed linking national satellite accounts on the stock of intangible assets with trade data would provide better understanding of the location of creation and uses of IP.

Potential Actions of National Tax Administrations 1.

2. 3.

closer cooperation with national statisticians and qualified academic researchers, including potential agreements to share tax return information on MNEs, as appropriate analysis and aggregated publication of future country-by-country reporting data of MNEs analysis and publication of aggregate tax return information from MNEs and domestic only firms, and details on income and expenses from intellectual property.

Potential Actions of National and International Policymakers and Analysts 1.

2.

3. 4.

recognize the importance of empirical data on cross-border IP flows and support adequate resources for the collection and analysis of such data increase transparency of tax incentives, particularly for the use of IP transactions by firms, such as declaration of the number of patents for which incentives are claimed increase transparency of IP ownership, which could improve markets for trading of IP and more consistent valuations reduce incentives (both tax and non-tax) for profit-shifting through changes to national rules and as part of international agreements



5.

6. 7.

    - 

adhere to the BEPS Inclusive Framework minimum standards on harmful tax practices, increased transfer pricing documentation, and revised transfer pricing guidelines support increased revenue mobilization in low-income countries through training and collaborative tax administration initiatives support improved dispute resolution mechanisms to resolve potential double tax situations arising out of multiple countries claiming taxing rights to the same cross-border income with subsequent adjustments to national statistics.

Academics 1. 2.

3.

increase focus on profit-shifting in low-income countries, including IP cross-border flows increase analysis of governmental incentives for IP and their potential for distortions of reported cross-border IP flows, including improved measures of embedded IP increase analysis of factors contributing to affiliate profitability. If imputations are needed, what are appropriate estimation methodologies of affiliate profitability?

Conclusions As described in this chapter, intangible assets play a critical role in how the modern economy operates with global fragmented production and associated trade in intermediate products. Understanding cross-border flows of IP knowledge flows is essential to analysing how modern economies operate. This in turn calls for better metrics of charges for the use of IP, since they are distorted by several factors, particularly MNE tax planning to reduce their global tax liabilities. Cross-border IP-related transfer mispricing and other tax minimization strategies distort current trade data, including cross-border IP payments, total imports and exports, GDP and productivity of individual countries. The empirical literature on profit-shifting, plus new analysis of CUIP flows, show their sensitivity to tax rates. The general magnitude of the measurement distortions between countries due to tax-induced profit-shifting is conservatively estimated at 35 per cent of total Charges for the Use of Intellectual Property. These distortions have been increasing over time.

               



Recent global initiatives, such as the OECD/G20 BEPS project and the recommendations suggested in this chapter will hopefully reduce taxinduced profit-shifting and improve global and national measurement of IP. Better national statistics on cross-border IP flows will help empiricalbased policymaking.

Bibliography Arora, Ashish, Fusfuri, A. and Gambardella, A., ‘Markets for Technology and their Implications for Corporate Strategy’ (2001), Industrial and Corporate Change 10, 2: 419–451. Athreye, S. and Yang, Y., ‘Disembodied Knowledge Flows in the World Economy’, Background Report to the WIPO 2011 World Intellectual Property Report. Published as WIPO Economic Research Working Papers, No. 3, Geneva: WIPO, 2012. Beer, Sebastian, de Mooij, Ruud A. and Liu, Li, ‘International Corporate Tax Avoidance: A Review of Channels, Magnitudes and Blind Spots’, IMF Working Paper No. 18/168, 2018. Clausing, K. A., ‘Tax-Motivated Transfer Pricing and US Intrafirm Trade Prices’ (2003), Journal of Public Economics 87, 9–10: 2207–2223. Clausing, K. A., ‘The Effect of Profit-Shifting on the Corporate Tax Based in the United States and Beyond’ (2016), National Tax Journal 69, 4: 905–934. Cobham, Alex and Jansky, Petr, ‘Global distribution of revenue loss from corporate tax avoidance: reestimation and country results’ (2018), Journal of International Development 30, 2: 206–232. Cristea, Anca and Nguyen, Daniel, ‘Transfer Pricing by Multinational Firms: New Evidence from Foreign Firm Ownerships’ (2016), American Economic Journal: Economic Policy 8, 3: 170–202. Crivelli, Ernesto, Keen, Michael and de Mooij, Ruud A., ‘Base Erosion, ProfitShifting, and Developing Countries’, Oxford University Centre for Business Taxation Working Paper WP 15/09, 2015. De Haan, M., Connolly, M., Peltola, R. and Dimova, T., ‘Measuring Global Production in the National Accounts and the Balance of Payments’, 24–30 August 2014, Paper Prepared for the IARIW 33rd General Conference, Rotterdam, the Netherlands. Dischinger, M. and Riedel, N., ‘Corporate Taxes and the Location of Intangible Assets within Multinational Firms’ (2011), Journal of Public Economics 95: 691–707. Dowd, Tim, Landefeld, Paul and Moore, Anne, ‘Profit-Shifting of U.S. Multinationals’ (2017), Journal of Public Economics 148: 1–13.



    - 

Dudar, Olena and Voget, Johannes, ‘Corporate Taxation and Location of Intangible Assets: Patents vs Trademarks’, No. 16-015, ZEW Discussion Papers, ZEW – Center for European Economic Research, 2016. European Commission, ‘Patent Boxes Design, Patents Location and local R&D’, IPTS Working Papers on Corporate R&D and Innovation, No. 6/2015, 2015. Evers, L., Miller, H. and Spengel, C., ‘Intellectual Property Box Regimes: Effective Tax Rates and Tax Policy Considerations’ (2013), ZEW Discussion Papers No. 13-070. Fetzer, James J., Mataloni, Jr., Raymond J. and Thompson, Sarahelen, ‘BEA’s Initiative to Expand and Reconcile Trade in Services Statistics: New Detail for Improved Analysis’, mimeo, 11 April 2017, available at www.gtap.agecon .purdue.edu/resources/download/8436.pdf. Griffith, Rachel, Miller, Helen and O’Connell, Martin, ‘Ownership of Intellectual Property and Corporate Taxation’ (2014), Journal of Public Economics 112: 12–23. Grubert, Harry, ‘Intangible Income, Intercompany Transactions, Income Shifting, and the Choice of Location’ (2003), National Tax Journal 56, 1: 221–242. Guvenen, Fatih, Mataloni, Jr., Raymond J., Rassier, Dylan G. and Ruhl, Kim J., ‘Offshore Profit-shifting and Domestic Productivity Measurement’, National Bureau of Economic Research Working Paper No. 23324, 2017. Hebous, Shafik and Johannesen, Niels, ‘At Your Service! The Role of Tax Havens in International Trade with Services’, CESifo Working Paper Series No. 5414, CESifo Group Munich, 2015. IMF, Balance of Payments and International Investment Position Manual, Sixth Edition (BPM6), Washington, DC: International Monetary Fund, 2009. IMF, Balance of Payments Statistics Yearbook and data files, available at https:// databank.worldbank.org/metadataglossary/africa-development-indicators/ series/BX.GSR.ROYL.CD, n.d. IMF, ‘Changing Patterns of Global Trade’, (Strategy, Policy, and Review Department Paper), 2012, available at www.imf.org/external/pubs/ft/dp/ 2012/dp1201.pdf. IMF, ‘Spillovers in International Corporate Taxation’, IMF Policy paper, Washington, DC: IMF, 2014. Karkinsky, T. and Riedel, N., ‘Corporate Taxation and the Location of Patents within Multinational Firms’ (2012), Journal of International Economics 88, 1: 176–185. Liu, Li, Schmidt-Eisenlhor, Tim and Guo, Dongxian, ‘International Tax Avoidance and Transfer Mispricing: Evidence from Linked Tax-Trade Statistics in the UK’, Oxford Centre for Business Taxation Working Paper No. 17/02, 2017. Madeuf, B. ‘International Technology Transfers and International Technology Payments: Definitions, Measurement and Firms’ Behaviour’ (1984), Research Policy 13, 3: 125–140.

               



OECD, Proposed Standard Method of Compiling and Interpreting the Technology Balance of Payments Data (TBP Manual), Paris: OECD Publishing, 1990. OECD, Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations, Organisation For Economic Co-operation And Development, Paris: OECD Publishing, 1995. OECD, Measuring Globalisation: OECD Handbook on Economic Globalisation Indicators 2005, Paris: OECD Publishing, 2005. OECD, Handbook on Deriving Capital Measures of Intellectual Property Products, Paris: OECD Publishing, 2014. OECD, Measuring and Monitoring BEPS, Action 11 – 2015 Final Report, OECD/ G20 Base Erosion and Profit-shifting Project, Paris: OECD Publishing, 2015, available at http://dx.doi.org/10.1787/9789264241343-en. Rassier, Dylan G., ‘Characteristics of Special Purpose Entities in Measures of U.S. Direct Investment Abroad’, 2015, available at www.bea.gov/papers/pdf/char acteristics-of-special-purpose-entities-in-measures-of-us-direct-investmentabroad.pdf. Torslov, Thomas, Wier, Ludvig and Zucman, Gabriel, ‘The Missing Profits of Nations’, NBER Working Paper No. W24701, 2018. United Nations Economic Commission for Europe (UNECE), Guide to Measuring Global Production, Geneva: United Nations Economic Commission for Europe (in particular, Chapter 7 on International transactions in intellectual property products), 2015, available at www.unece.org/fileadmin/DAM/stats/ publications/2015/Guide_to_Measuring_Global_Production__2015_.pdf. UN, IMF, OECD, Eurostat, UNCTAD, UNWTO and WTO, Manual on Statistics of International Trade in Services 2010 (MSITS2010), Geneva, Luxembourg, Madrid, New York, Paris and Washington, DC: United Nations, IMF, OECD, Statistical Office of the European Union, United Nations Conference on Trade and Development, World Tourism Organization and WTO, 2011. UNCTAD, World Investment Report 2015: Reforming international investment governance, UN Reference: UNCTAD/WIR/2015, 2015b. WIPO, ‘The Changing Nature of Innovation and Intellectual Property, Chapter 1’, in World IP Report 2011, Geneva: WIPO Economics and Statistics Division, 2011, pp. 23–72. WIPO, World Intellectual Property Report 2017: Intangible Capital in Global Value Chains. Geneva: World Intellectual Property Organization, 2017.

7 Global Ebbs and Flows of Patent Knowledge       .         ,     .         .       

Abstract We use patent citation data to map the flows of technical knowledge, and in particular to test three hypotheses. First, that the net flow of technical knowledge tends to be from more developed to less developed countries. Second, that the trajectory of climate change-mitigation knowledge flow increased (especially from developed to developing countries) after the 1997 Kyoto Protocol and the 2015 Paris Agreement on climate change. Third, that the trajectory of pharmaceutical knowledge flow increased after the conclusion of the TRIPS Agreement. We find that massive amounts of technical information are exchanged, among countries around the world, in sometimes surprising patterns. Developed country patents supply a disproportionate amount of patent citations compared to patents from developing countries, and the importance of technical knowledge thereby transferred is disproportionately high. However, we observe some evidence that this imbalance in knowledge flow may be beginning to reverse itself, as developing country patents have become increasingly cited by those in developed countries, and the importance of developing world patents has grown. We did not discern evidence that the Kyoto Protocol has spurred knowledge flow in climate-mitigation technology, but found some preliminary evidence consistent with an effect by the Paris Agreement. We did find evidence consistent with the possibility that the TRIPS Agreement may have spurred pharmaceutical (and more general) knowledge flow.

Introduction Patents and patent applications contain rich sources of scientific and technical information. By their very name, patents – derived from a Latin word meaning ‘open’ – make their information freely available to the public. They describe new devices, machines, systems, molecules and even living organisms, as well as methods for making and using them. Transfer of technical knowledge among people, firms, institutions and governments 

       



has long relied upon the information patent documents (hereafter simply ‘patents’) disclose. One may take advantage of a feature of patent law to track the flow of technical ideas among patents by constructing a network from the citations that patents make to and from each other. The result is a patent citation network. The worldwide patent citation network is immense, with almost half a billion citations among more than 100 million patents. In this chapter, we use the worldwide patent citation network to infer the flow of technical information among countries. Technical knowledge flows among inventors, organizations (e.g. firms, research institutes, universities) and countries. Patent citation data can be used to estimate these flows. Here we focus on knowledge flow among countries or other jurisdictions, calculating pairwise citations – both raw and importance-weighted – to and from them. This method offers new insights into global patterns of technical knowledge flow. Our study suggests that the use of big patent data and hierarchical mapping network approaches may offer new, and potentially useful, insights into the flow of technical knowledge among countries and other patent jurisdictions, within specific technologies, and over defined periods of time. Scientific and technical publications, and the citations among them, may be envisioned as forming a vast network. In this network, publications are ‘nodes’, while citations among nodes are ‘links’. One source of such publications is patents and patent applications. The architecture of such a patent network represents millions of individual choices inventors, and patent agents or attorneys who represent them, have made about how to situate inventors’ new ideas within the wider context of existing knowledge. If the computationally-challenging task of constructing a patent citation network is done successfully, the resulting lattice of interlinked inventions may hold within its structure immense amounts of valuable information concerning the evolution of, and relationships between, technologies. There are myriad approaches to building a patent citation network. We use eigenvector centrality and hierarchical graphical approaches to construct comprehensive citation networks. This results in several salutary features useful for analysing not only network structure at defined moments in time but also changes in network structure over time. The network contains millions of ‘natural technology clusters’ grouped together by their citations, with these natural technology clusters running the gamut of granularity from the most specific technology areas to the most inclusive areas. In part by virtue of the nested organization of its natural technology clusters, the structure of the patent citation network holds a wealth of



     .         .

information about where knowledge is generated, where it flows and how patterns have already, and likely shall in the future, change. In addition to elucidating general patterns of technical knowledge flow, we also test three specific hypotheses designed to shed light on flow of technological information among countries. The first is general, and is inspired by a growing interest in technology transfer between the developed and developing world: 1.

The net flow of technical knowledge tends to be from more developed to less developed countries.

The other two hypotheses investigate two particular fields of technology – climate-change mitigation technology (with wind power technology as a proxy) and pharmaceutical technology – which several prominent international agreements (e.g. the 1995 World Trade Organization (‘WTO’)1 Agreement on Trade-Related Aspects of Intellectual Property Rights (‘TRIPS’))2 (which, inter alia, mandated protection for pharmaceutical technology), and the Framework Convention on Climate Change’s Kyoto Protocol and Paris Agreement (dealing, inter alia, with climate-change mitigation technology) were expected by many to encourage and increase technical knowledge transfer, especially from developed countries to developing countries. These more specific hypotheses are: 2.

3.

The trajectory of climate-change mitigation knowledge flow increased (especially from developed to developing countries) after the 1997 Kyoto Protocol3 and the 2015 Paris Agreement4 on climate change; and The trajectory of pharmaceutical knowledge flow increased after the conclusion of the TRIPS Agreement.

We find that massive amounts of technical information are exchanged, among countries around the world, in sometimes surprising patterns. 1

2

3

4

The WTO is an international organization dedicated to the promotion and administration of trade among countries around the world. (See www.wto.org/index.htm.) Administered by the WTO, the TRIPS is a multilateral trade agreement that attempts to standardize intellectual property rights among WTO members. (See www.wto.org/english/ tratop_e/trips_e/trips_e.htm.) The Kyoto Protocol is a 1997 treaty intended to implement the 1992 United National Framework Convention on Climate Change. (See https://unfccc.int/process-and-meet ings/the-kyoto-protocol/what-is-the-kyoto-protocol/what-is-the-kyoto-protocol.) The Paris Convention is a 2015 treaty intended to implement the 1992 United National Framework Convention on Climate Change. (See https://unfccc.int/process-and-meet ings/the-paris-agreement/the-paris-agreement.)

      



Background Overview In this study, we focus on an important subset of the worldwide network of scientific and technical knowledge composed of granted patents and patent applications (collectively referred to as ‘patents’ unless otherwise noted). The eigenvector centrality approach we use to construct and analyse the worldwide patent citation network results in estimates for each patent of its ‘importance’ within the context of the network. The dataset on which we primarily rely here is the ‘PATSTAT’ database compiled, annotated and maintained by the European Patent Office (‘EPO’).5 The importance approach can be used to study not only the worldwide patent system, but also national and regional patent systems. It also allows comparisons between jurisdictions of choice (e.g. the United Kingdom and Japan, the United States and the European Patent system6).

Previous Research We published the first patent citation network study to use the importance approach in 2017.7 Entitled All Patents Great and Small: A Big Data Network Approach to Valuation (2017) this project focused on the United States (‘US’) patent system. Facilitated by availability of highquality, weekly-updated patent data for bulk downloading, this research project offered confirmation that the eigenvector centrality approach works well for patent citation networks. However, the US patent system is only a subset of the worldwide patent system. For this chapter, we constructed a global patent citation network that incorporates not only US patents, but all patents from other countries and jurisdictions located within the PATSTAT database.

5

6

7

In the past, we have also used bulk patent data provided by the United States Patent and Trademark Office (‘USPTO’). See Andrew W. Torrance and Jevin D. West, ‘All Patents Great and Small: A Big Data Network Approach to Valuation’, 20 Virginia Journal of Law & Technology 466 (2017). Here ‘European Patent system’ refers to the regional patenting authority governed by the European Patent Convention (EPC). See Andrew W. Torrance and Jevin D. West, ‘All Patents Great and Small: A Big Data Network Approach to Valuation’, 20 Virginia Journal of Law & Technology 466 (2017).



     .         .

International Climate Agreements The UN Framework Convention on Climate Change (‘UNFCCC’) has given rise to two major implementing agreements: the Kyoto Protocol (1997) and the Paris Agreement (2015). The time period covered by our data allows some consideration of effects these treaties may have had in knowledge transfer in climate-mitigation technology, represented here by wind power technology. The main technology transfer provision in the Kyoto Protocol is found in Article 10(c), which requires countries to ‘promote, facilitate and finance, as appropriate, the transfer of, or access to, environmentally sound technologies, know-how, practices and processes pertinent to climate change, in particular to developing countries, including the formulation of policies and programmes for the effective transfer of environmentally sound technologies that are publicly owned or in the public domain’.8 Notably, Article 10(c) also encourages countries to develop programmes that ‘create an enabling environment’ for the private sector to transfer environmentally sound technologies to developing countries.9 This provision, however, does not require participation by the private sector. The Kyoto Protocol also imposes a duty on the developed countries to the Agreement to establish policies for technology transfer that minimize adverse economic, social and environmental impacts on developing countries.10 The Kyoto Protocol as a whole has received criticism for not effectively reducing greenhouse gas emissions, even by the former executive secretary of the UN Climate Secretariat, Yvo de Boer.11 Despite this criticism, de Boer has suggested that the Kyoto Protocol has achieved success by laying the foundation for technology transfer of environmentally sound technologies.12 About a third of all projects related to the Clean Development Mechanism13 involve some sort of technology

8

9 10 11

12 13

Kyoto Protocol to the United Nations Framework Convention on Climate Change Article 10(c), 11 December 1997, UN Doc FCCC/CP/1997/7/Add.1, 37 ILM 22 (1998) (hereinafter Kyoto Protocol). Id. Kyoto Protocol, supra n. 8, Article 4, ¶ 14. David Adam, ‘Analysis: Has the Kyoto Protocol Worked?’, The Guardian (7 December 2008, 7:01 PM EST), available at www.theguardian.com/environment/2008/dec/08/ kyoto-poznan-environment-emissions-carbon (‘Has Kyoto worked? “In terms of emission reductions achieved, the answer would be no”, De Boer said.’). Id. Kyoto Protocol, supra n. 8, Article 12.

      



transfer from developed to developing countries.14 Moreover, projects that involved some form of technology transfer under the Kyoto Protocol have accounted for 59 per cent of estimated annual emissions reductions, suggesting that projects involving technology transfer may indeed have larger effects than those projects lacking technology transfer.15 The rate of technology transfer does differ among types of projects; projects related to energy distribution, hydro, cement, fugitive emissions, transport and energy efficiency have low rates of technology transfer, ranging from 0 to 27 per cent, while projects involving tidal power, CO2 capture, agriculture, HFCs and energy-efficient households have much higher technology transfer rates, ranging from 56 to 100 per cent.16 The types of technology transfer (transfer of equipment, knowledge, or both) also differ between project types.17 In 2015, most of the world’s countries agreed to the Paris Agreement, an implementation treaty intended to update the Kyoto Protocol. Article 10 of the Paris Agreement sets out goals related to technology transfer by using the ‘Technology Mechanism’ created by the UNFCCC.18 The Paris Agreement reinforces the importance of technology transfer in the effort to reduce greenhouse gas emissions and combat climate change,19 requires parties to the agreement to ‘strengthen cooperative action on technology development and transfer’,20 and requires parties to provide financial support for the facilitation of access to technology to developing countries.21 In Article 11, the Agreement emphasizes the use of technology transfer to assist developing countries in capacity building.22 The Agreement also contains accountability mechanisms wherein developing countries must report information related to the technology received through technology transfer

14

15

16 17 18

19 20 21 22

Zhong Fa Ma, ‘The Effectiveness of Kyoto Protocol and the Legal Institution for International Technology Transfer’, 37 Journal of Technology Transfer 75, 83 (2010). Stephen Seres, Analysis of Technology Transfer in CDM Projects 7 (2008), available at http://cdm.unfccc.int/Reference/Reports/TTreport/TTrep08.pdf. Id. at 24. Id. at 13. UN Framework Convention on Climate Change, Adoption of the Paris Agreement, Article 10, 12 December 2016, UN Doc. FCCC/CP/2015/10/Add.1, available at http://unfccc.int/ files/home/application/pdf/paris_agreement.pdf. (hereinafter Paris Agreement). Id. at Article 10 ¶ 1. Id. at Article 10 ¶ 2. Id. at Article 10 ¶ 5. Id. at Article 11 ¶ 1.



     .         .

and developed countries must report information related to technology they have transferred to developing countries.23

Methods In this chapter, we provide only a brief summary of the methods by which we construct and analyse the worldwide patent citation network we use to estimate technical knowledge flow. A detailed explanation of these methods is available in our 2017 article, All Patents Great and Small: A Big Data Network Approach to Valuation.24

Network Analysis Patents often cite other patent documents. Over time these citations accumulate.25 A patent citation network spanning all patent systems in the world is a valuable tool for studying worldwide patterns of scientific information flow dynamics, including those elucidating technological innovation, patenting effort, and national and international patent law regimes. Major technical barriers have hampered efforts to calculate a worldwide patent citation network. We present a powerful new approach to patent citation network calculation based on big data and network analytics. In our 2017 article, All Patents Great and Small: A Big Data Network Approach to Valuation, we provided several lines of evidence demonstrating the accuracy and explanatory power of our approach as applied to the US patent corpus.26 For this chapter, we applied these methods to the construction of a worldwide patent citation network, encompassing the patent systems of all major countries and jurisdictions, based on the fall 2018 PATSTAT bibliographic data that include more than 112,331,073 million patent documents connected by 265,303,432 patent citations. We provide a brief description of the structure of this network, and then highlight novel insights it provides about the worldwide evolution of technological fields and comparative differences in 23 24

25

26

Id. at Article 13 ¶¶ 9, 10. See Andrew W. Torrance and Jevin D. West, ‘All Patents Great and Small: A Big Data Network Approach to Valuation’, 20 Victoria Journal of Law & Technology 466 (2017). Because citations accumulate over time, all other factors being equal, older patents tend to have attracted more citations than more recent ones. A consequence of this pattern of accumulation over time is that metrics based on citations tend to be biased downwards as time approaches the present. See Andrew W. Torrance and Jevin D. West, ‘All Patents Great and Small: A Big Data Network Approach to Valuation’, 20 Victoria Journal of Law & Technology 466 (2017).

      



patenting behaviour among countries and regions. In addition, we present an overview of original network analytics approaches to analysing large datasets. These offer promising new, powerful and efficient approaches for constructing and analysing complex networks like the worldwide patent citation network.

Patent Citations Citations to and from patents tend to be indicators of both private value to their owners and social value to society more generally.27 Patent citations have been widely used in patent valuation analysis.28 They can be informative about firm value,29 useful in assisting universities to predict which of the patents they own will most likely be licensed30 and indicative of whether a patent application will be granted.31 Patent citations have been found to correlate well with likelihood of litigation32 and patent value.33 Beyond economic value alone, forward citations can provide good estimates of the technological importance of inventions disclosed in patents.34 27

28

29

30

31

32

33

34

Manuel Trajtenberg, ‘A Penny for Your Quotes: Patent Citation and the Value of Innovations’, 21 Rand Journal of Economics 172 (1990). Bronwyn H. Hall, Adam B. Jaffe and Manuel Trajtenberg, Market Value and Patent Citations: A First Look (National Bureau of Economic Research, Working Paper No. 7741, 2000); Dietmar Harhoff, Katrin Cremers, Francis Nairn, Frederic M. Scherer and Katrin Vopel, ‘Citation Frequency and the Value of Patented Inventions’, 81 Review of Economics and Statistics 511, 511 (1999); Jean O. Lanjouw and Mark Schankerman, ‘Characteristics of Patent Litigation: A Window on Competition’, 32 Rand Journal of Economics 129 (2001). Bronwyn H. Hall, Adam Jaffe & Manuel Trajtenberg, ‘Market Value and Patent Citations’, 36 Rand Journal of Economics 16 (2005). Bhaven N. Sampat and Arvids A. Ziedonis, ‘Patent Citations and the Economic Value of Patents’, in H. F. Moed, W. Glänzel and U. Schmoch (eds), Handbook of Quantitative Science and Technology Research, p. 277. Alfons Palangkaraya, Elizabeth Webster & Paul H. Jensen, ‘Misclassification Between Patent Offices: Evidence From a Matched Sample of Patent Applications’, 93 Rev. Econ. & Stat. 1063 (2011). Jean O. Lanjouw and Mark Schankerman (2003) ‘Enforcing Patent Rights: An Empirical Study’, in Empirical Economics of Innovation and Patenting, 2003-03-14–2003-03-15, p. 14, available at http://eprints.lse.ac.uk/58574/ (‘The probability of litigation increases with . . . forward citations per claim . . . the likelihood of a suit falls with the number of backward citations per claim.’) John R. Allison, Mark A. Lemley, Kimberly A. Morre and R. Derek Trunkey, ‘Valuable Patents’, 92 The Georgetown Law Journal 435, 451 (2004). Mark P. Carpenter, Francis Narin and Patricia Woolf, ‘Citation Rates to Technologically Important Patents’, 3 World Patent Information 160 (1991); Francis Narin, Elliot Noma and Ross Perry, ‘Patents as Indicators of Corporate Technological Strength’, 16 Res. Pol’y



     .         .

One limitation of our approach is that it is very difficult to parse inventions originating in a country from those originating elsewhere. To illustrate, imagine all of the patent applications were filed in Nigeria. Some of these patent applications claim technologies invented by locals; however, other Nigerian patent applications claim inventions that arose in other countries, but were allowed to be filed as Nigerian patent applications under the Paris Convention or the Patent Cooperation Treaty. Both of these categories of patents count as Nigerian for the purposes of our analyses. In the future, more granular data may allow these two types of patents to be distinguished easily, at which time more detailed analysis will be possible. There may also be bias in the number of developing world patent citations due to institutional limitations in the developing world. It requires considerable capacity, both technological and human capital, to collect and search prior art from other countries, leading, perhaps, to more patent citations in patents filed in the developed world than in the developing world. Our methods minimize this bias by focusing more on patent importance than on raw patent citations; this reduces the influence of large numbers of patent citations and of citations computed on a per patent basis because quality of citation tends to trump quantity of citation. An approach that could be used in the future would be to increase the size of the analytical unit used to construct patent citations networks to patent families rather than relying on individual patents and patent applications. Such an approach might help to dispel the illusion that, when large US, European and Japanese companies, such as those in the pharmaceutical industry, increasingly choose to file foreign versions of their patents in developing world countries (e.g. China, India, Brazil), this necessarily indicates increases in technological advancement there.

An ‘Importance’ Approach The most powerful way currently known to weight citations accurately is to construct a patent citation network in which the position of each patent helps determine its value. To analyse our patent network, we use

143 (1987); Michael B. Albert, D. Avery, Francis Narin and Paul R. McAllister, ‘Direct Validation of Citation Counts as Indicators of Industrially Important Patents’, 20 Res. Pol’y 251 (1991).

      



the Eigenfactor metrics35, which have become the gold standard in ranking scholarly journals.36 The algorithm captures a random walker following hyperlinks (links) from webpage (nodes) to webpage. The Eigenfactor algorithm captures a random process on scholarly citation networks.37 For patent citation networks, we use a modified version of the Eigenfactor algorithm called the article-level Eigenfactor (ALEF).38 To calculate the ALEF scores, we use methods described by West and Vilhena (2014)39 and West et al. (2016)40 By sewing together the citations from the countries represented in PATSTAT, we offer a better understanding of how patents organize by country and technology areas at very fine, subcommunity scales. This complements the humancurated, patent classification systems and provides a different lens for how to evaluate patent influence and technology trends.

Patent Networks and Knowledge Flow This particular focus of this study involves a comprehensive analysis of the flow of technical knowledge within the worldwide patent system.41 It does so using two metrics. The first of these is based upon raw citations, 35

36

37

38

39

40 41

Jevin D. West, Theodore C. Bergstrom and Carl T. Bergstrom, ‘The Eigenfactor Metrics: A Network Approach to Assessing Scholarly Journals’, 71 College & Research Libraries 236 (2010). They are now included in Thomson-Reuters’ Journal Citation Reports (JCR). The underlying algorithm is similar to the PageRank algorithm developed by the founders of Google, Larry Page and Sergey Brin. Lawrence Page, Sergey Brin, Rajeev Motwani and Terry Winograd, ‘The PageRank Citation Ranking: Bringing Order to the Web, Stanford InfoLab’, 1999, available at http://ilpubs.stanford.edu:8090/422/1/1999-66.pdf. The algorithm was placed first in North America and second worldwide in Microsoft Research’s WSDM Cup Challenge, a 2015 contest whose goal was to statically rank tens of millions of articles from the scholarly literature. The contest provided additional evidence of the advantages of using the network rather than just counting raw citations. Microsoft Research, WSDM Cup Challenge, 2015, available at www.microsoft.com/enus/research/publication/wsdm-cup-2016-entity-ranking-challenge/. Jevin D. West, Ian Wesley Smith and Carl T. Bergstrom, ‘A Recommendation System Based on Hierarchical Clustering of an Article-Level Citation Network’, 2 IEEE Transactions on Big Data 113 (2016). Jevin D. West and Daril A. Vilhena, ‘A Network Approach to Scholarly Evaluation’, in Blaise Cronin and Cassidy R. Sugimoto (eds), Bibliometrics & Beyond: Metrics-Based Evaluation, 2014, pp. 151–166. See US Patent No. 9,589,051, at [0032] (filed 1 February 2013). We do not purport to demonstrate herein that the technical information disclosed in patent documents is meaningful. Instead, we simply show the flows of such information via citations, both raw and weighted.



     .         .

which is the number of times one patent cites a different patent. The second is importance, which reflects not simply the number of citations, but the relative influence of each citation. For every available national (e.g. New Zealand, Italy, Brazil), regional (e.g. the European Patent Convention), and international (e.g. Patent Cooperation Treaty) patent jurisdiction, we enumerated the citations to and citations from all national and jurisdictional pairs as well as the importance of these citations.

Data Based on the United Nations classification42, we divided countries and jurisdictions into two categories: developed and developing. In addition, we grouped patents originating under the jurisdiction of the European Patent Convention (‘EPC’) as developed. Turkey and Serbia are both usually classified as developing countries despite being members of the EPC; however, because their contributions to the EPC are still quite minor, their inclusion is unlikely to affect the results. It is difficult to categorize climate-change mitigation technology because it encompasses a diverse array of more internally-consistent technological fields, such as solar power, carbon sequestration, energy efficiency, hydroelectric power, tidal power, power system design, high-efficiency engines, power management software and wind power. Wind power technology is the largest and most coherent of these fields. Consequently, we used wind power technology, a cluster of about 33,186 patents, as a proxy to represent all climate-change mitigation technologies.43 It is comparatively easier to classify pharmaceutical technology because of its relatively high internal coherence as a technological field. We identified a large natural cluster having 4,526,982 patents that includes a large majority of patents covering pharmaceutical technologies.44

Caveat Regarding Recent Patent Data For all figures showing filing, citation and importance trends for pharmaceuticals or climate-mitigation technologies, the decay in values that occurs towards the right-hand side of the graphs is a data artefact. Because most 42

43 44

See Table A, available at www.un.org/en/development/desa/policy/wesp/wesp_current/ 2014wesp_country_classification.pdf. These patents are identified in our network as Cluster 1: 168. These patents are identified in our network as Cluster 3.

       



patents experience a lag in publication, recent data tends to underestimate total numbers. Both citations and importance accumulate over time, meaning that, as measurements of these approach the present day, they will tend to decrease in magnitude. Thus, we tend to avoid making conclusions about these data for periods of time close to the present day.

Results We derive estimates of technical knowledge flow from the structure and dynamics of the worldwide citation network. The information disclosed in patents certainly does not represent the sum total of technical knowledge. Important alternative information sources include the scientific literature, webpages and data available on the Internet, and knowledge residing within the heads of scientists and engineers. Nevertheless, patents are indeed one of the major sources of technical knowledge. We examine worldwide technical knowledge flow, but also flow between developed and developing countries, in general, and for the specific technological fields of wind power and pharmaceuticals. For the latter, we estimate flow of technical knowledge of all kinds, as well as flows in the specific technological fields of wind power (as a proxy for climatechange mitigation technologies) and pharmaceuticals and biotechnology. To illustrate our methods and data, we begin with the pairwise technical knowledge flow between the US and China.

An Illustrative Example: China The pattern of citations between China and the US provides an illustration of our results based on both raw citations and citations weighted for importance. Consider the citation patterns between Chinese and US patent documents. Chinese patent documents have cited US patents or patent applications on 2,209,396 occasions, while US patent documents have cited Chinese documents only 592,154 times. Thus, Chinese patent documents cite US patent documents 3.73 times more often than US patent documents cite Chinese patent documents. The second metric is importance. Chinese patents have cited US patents that have a mean importance of 6.91. By comparison, US patents have cited Chinese patents that have a mean importance of 1.46. This yields an importance ratio of about 4.73:1 in favour of US patents. More broadly, the mean importance of all US-cited patents is 21.63 compared



     .         .

to a mean importance of 0.84 for all cited Chinese patents. This yields an importance of ratio of about 25.75:1, again in favor of US patents. One interpretation of these results is that US technical knowledge embodied within patents has flowed much more robustly towards, rather than from, China. Furthermore, measurements of knowledge flow derived from raw patent citations substantially underestimate the magnitude of this knowledge flow.

Trends in Patent Numbers The rise in patents and patent applications over the past half century has been prodigious (Figure 7.1). Since 1970, they have increased from just under 500,000 to a high point of more than 5,500,000 in 2017. During the same time period, the overall increase in US patent numbers has been from about 65,000 to over 725,000. Both US and worldwide patent numbers have swelled 11-fold in this interval.

Figure 7.1

Trends in patent numbers

Source: Authors’ analysis.

      



Numbers of both worldwide and US patents declined substantially after 2017. This fall is almost certainly an artefact caused by incomplete data. The fall 2018 PATSTAT dataset analysed for this chapter was released in October 2018, so it lacks patent data for at least the final quarter of 2018. In addition, there are delays by many national and international patent authorities in providing the most recent patent data to the European Patent Office for incorporation into PATSTAT. Accurate patent numbers for 2018 will become clearer with the passage of time. In the meantime, we suggest that these most recent numbers be regarded with scepticism.

Trends in Interjurisdictional Patent Citations Over the 50-year period starting in 1970, there are noticeable patterns in the proportion of patent citations made to patents within the same jurisdiction (‘intrajurisdictional’ citation) and those made to patents in other jurisdictions (‘interjurisdictional’ citation) (Figure 7.2). This pattern provides a window on the flows of technical knowledge among countries. It also sets a baseline for comparing trends in more specific technological fields, like wind power and pharmaceuticals. Figure 7.2 depicts an initial interjurisdictional citation rate of about 13 per cent in 1970. This modest level remained within a narrow range until it

Figure 7.2 Rates of interjurisdictional citation Source: Authors’ analysis.



     .         .

plummeted to about 6 per cent in 1975, after which it grew rapidly, more than quadrupling to about 27 per cent in 1982. After several years of relative stability, the rate began rising again in 2008, growing to about 38 per cent in 1997, fluctuating little until 2001, when it began to plummet, reaching about 18 per cent in 2005. It changed little until 2010, after which it rose sharply to about 29 per cent in 2012, moving little through 2018. Three broad, but contrasting, changes in rates of knowledge flow emerge from Figure 7.2. The first is the remarkable decades-long increase in rates of interjurisdictional citation from 1975 to 1997, during which technical knowledge flow increased more than six-fold. This was reversed by a precipitous fall of more than 50 per cent between 2000 and 2005. Finally, interjurisdictional citations rose suddenly by about 45 per cent between 2010 to 2012. To generalize, technical knowledge flow was relatively stable, grew prodigiously, stabilized, fell rapidly, stabilized, grew rapidly again and then stabilized again.

Knowledge Flow between the Developing and Developed World A major focus of this chapter is technical knowledge flow between developed and developing countries. As noted in the above section on methodology,45 we have employed a United Nations taxonomy that divides countries around the world into the categories of developed and developing46. In our analysis, there are 38 developed countries and jurisdictions (including the EPC area) and 198 developing countries. A vital principle for interpreting the patent citation data presented below is that, as one approaches the current day (2019 at the time of writing), deficiencies in data collected from patent offices around the world become more serious. Some patent offices report their data with alacrity, but others report theirs at a more sluggish pace.47 Consequently, citation data decrease in completeness as one approaches the present. Another interpretive principle concerns patent citation and importance data. As described in the above subsection on network analysis,48 citations naturally accumulate with the passage of time. Thus, a relative 45 46

47 48

P. 224 supra. See Table A, available at www.un.org/en/development/desa/policy/wesp/wesp_current/ 2014wesp_country_classification.pdf. Personal communications with EPO staff responsible for PATSTAT. P. 224 supra.

      



dearth of citations by more recent patents is to be expected. Because citations among patents in the worldwide patent citation network are the basis for calculating importance, both citation numbers and importance are impacted by recency; confidence in the accuracy of both increases as time before the present day increases.

Raw Citations The number of raw patent citations provides a straightforward metric of knowledge flow. From 1985 to 2005, there were very few patent citations between developed world patents and developing world patents (Figure 7.3). Nevertheless, until 2006, the number of citations from developed to developing countries was almost twice as large as in the other direction. After 2005, rates of citations in both directions rose markedly, but much more rapidly for citations from developing to developed country patents. In addition, the polarity of citations reversed in 2006, with many more citations thereafter from developing to developed country patents. Also depicted in Figure 7.3 are numbers of citations from US to developing world patents and from China to

Figure 7.3 Patent citations between developed and developing countries Source: Authors’ analysis.



Figure 7.4

     .         .

Importance of citing patents between developed and developing countries

Source: Authors’ analysis.

developed world patents. Since 2011, China has accounted for the majority of citations from developing to developed world patents, while the US accounted for the majority of citations from developed to developing world patents since at least 1985.

Citations Weighted by Importance As discussed in the above section on methodology,49 patent citations weighted by their importance in the worldwide patent citation network may provide a more accurate metric for technical knowledge flow than do raw patent citations. Figure 7.4 shows the aggregate importance of developed country patents citing developing country patents, and vice versa. In addition, it depicts US patents citing patents in the developing world and Chinese patents citing patents in the developed world. From 1985 to 2005, the total importance of developed world patents remained far greater than those in the developing world, with the former ranging from 19,545 to 57,701, and the latter from 0 to 21. Beginning in 2006, the importance of developing country patents became substantially greater, 49

P. 224 supra.

      



Figure 7.5 Importance of cited patents between developed and developing countries Source: Authors’ analysis.

almost equalling the importance of developed country patents by 2010, and subsequently dwarfing them until 2016. Figure 7.4 also illustrates that the patterns in cited patent importance in developed and developing countries are overwhelmingly accounted for by the contributions of US and Chinese patents. Figure 7.5 illustrates the aggregate importance of developed country patents cited by developing country patents, and vice versa. The pattern for cited patents is noticeably different from that for citing patents. The aggregate importance of developed world patents cited by developing world patents increased from 5,213 in 1985 to 15,332 in 1998, before subsiding gradually to 11,224 in 2005. Beginning in 2006, though, their total importance rose rapidly, almost 50-fold, to 524,657 in 2017. The trend was strikingly similar for US patents being cited by developing world patents, with US patents again accounting for most of the importance. The aggregate importance of developing world patents cited by developed world patents increased rapidly, from a negligible 1.14 in 1985 to 1672,98 only two years later, in 1987. After remaining relatively stable until 1991, their importance began to rise rapidly once again, reaching 32,020,25 in 1995. This reflected a rise in importance of more than 28,000 times in a decade. By 1998, their importance had fallen



     .         .

sharply by almost 10 times, to 3714,60. With several temporary interruptions, importance rose rapidly, reaching a high of 4893464,10 in 2013, before shallowly subsiding. Until 1992, the importance of these cited developing world patents lagged well behind that of cited developed world patents, but, from 1993 to 1996, the former surmounted the latter. After falling below cited developed world patents in importance again until 2000, cited developing world patents established and maintained a huge lead in importance over those from the developed world. The pattern for Chinese patents shows substantial differences. After some initial fluctuations at very low levels from 2002 to 2006, the importance of these Chinese patents began a remarkable rise, reaching above 3,000,000 by 2012, and accounting for a majority of developing world patent importance by 2011. To put this prodigious rise in perspective, from 2002 – the first year for which we have Chinese data – to 2014, the aggregate importance of Chinese patents cited by patents in the developed world increased by more than 180,000 times. An alternative method of measuring knowledge flows is to consider mean, rather than aggregate, importance. Figure 7.6 shows the mean importance of developed country patents citing developing country patents, and vice versa. The mean importance of citing developed world patents remains above 4.0 until 2000, after which it declines gradually. The mean importance of citing developed world patents is

Figure 7.6 countries

Mean importance of citing patents between developed and developing

Source: Authors’ analysis.

      



much greater than citing developing world patents throughout the entire time period from 1985 to 2018. In fact, until 2010 this difference is one or two orders of magnitude. From 2010 onwards, the disparity narrows somewhat. The trend for citing US patents parallels that for developed world patents, though the former is consistently substantially greater than the latter. The trend for citing developing world patents is distinctly dissimilar to that of citing developed world patents. From 1988 until 2004, the importance of citing developing world patents fluctuated near a value of zero, after which it increased rapidly, pausing for several years after 2006 before reaching a high of 0.29 in 2011, and, then, decreasing steadily thereafter. From 2010 until 2017, citing developing world patents maintained importances only several times smaller than those for citing developed world patents. For all years in which Chinese patent data are available, the mean importance of citing Chinese patents was greater than the importance of citing developing world patents. Figure 7.7 illustrates the mean importance of developed country patents cited by developing country patents, and vice versa. Except for the first two years of data, 1985 and 1986, developing world patents cited by developed world patents have been substantially more important than developed world patents cited by those in the developing world. From 1987 to 2018, this difference has remained at about half an order of

Figure 7.7 Mean importance of cited patents between developed and developing countries Source: Authors’ analysis.



     .         .

magnitude, and the mean importance of these patents has consistently been very high, ranging from a low of 3.41 in 2016 to a high of 14.20 in 2002. The mean importance of these patents demonstrated a robust growth trend from 1987 to 2002, remained steady until 2005, fell by more than 50 per cent in 2006 and thereafter stayed relatively flat through 2018. The mean importance of cited Chinese patents was substantially higher than cited developing world patents during the period in which Chinese data first became available, moved almost in lockstep with cited developing world patents from 2010 until 2016 and then fell substantially below the latter in 2017 and 2018. The trends in mean patent importance for both cited developing world and US patents showed much less fluctuation. Starting at around 1.0 in 1985, they rose gradually to new highs in 1997, then declined gently back down to about 1.0 in 2005, remained relatively close to that level until 2016 and, finally, increased rapidly in 2017 and 2018. Throughout almost the entire period from 1985 until 2018, the mean importance of cited developed patents was higher than cited US patents.

Knowledge Flow in Specific Technologies In addition to exploring general trends in technical knowledge flow, we also examine two particular technologies: wind power and pharmaceuticals. Figure 7.8 illustrates the trends in patents in these two technological fields over time, and compares them with trends for all patents.

Figure 7.8

Trends in total, wind power and pharmaceutical patents over time

Source: Authors’ analysis.

      



Pharmaceutical patent numbers comprise approximately 2.5 per cent and 5 per cent of all patents. Wind power patents lag far behind this amount, consistently making up less than 0.1 per cent of all patents. From 1970 until 2018, both total patents and pharmaceutical patents increased in numbers by about an order of magnitude. By contrast, wind power patents increased by about two orders of magnitude over the same period of time.

Wind Power Technology Figure 7.9 depicts the growth over time of wind power patents. Numbers of patents in this technology cluster remained quite modest for almost three decades from 1970, not rising above 100 until 1980, nor above 200 until 1999. From 1998 until 2006, numbers of patents grew more than 3.5-fold, from 183 to 656. Then, from 2006 to 2012, these numbers accelerated substantially to 33,134, an almost five-fold increase. As explained above, the decay in numbers since 2012 is difficult to interpret, likely reflecting not a decline in wind power patents, but, rather, incomplete reporting of recent data from national patent offices. Figure 7.10 shows the interjurisdictional patent citation rate between developed and developing countries for wind power technology. The number of patents in this technology field is a small subset of all patents across all technologies. Though characterized by more fluctuations, the pattern in Figure 7.10 is generally similar to that for all technological

Figure 7.9 Trend in wind power patents over time Source: Authors’ analysis.



Figure 7.10

     .         .

Rates of interjurisictional citations for wind technology

Source: Authors’ analysis.

Figure 7.11 countries

Wind power patent citations between developed and developing

Source: Authors’ analysis.

fields (Figure 7.2), and replicates the overall rise-fall-rise pattern, though its rate of interjurisdictional citation tends to shift higher throughout. Figure 7.11 illustrates trends in the aggregate citations between wind power patents originating in the developed and developing world. It also includes citations from US patents to those in the developing world as well as citations from Chinese patents to those in the developed world. There were very few patent citations between 1985 and 2006; in fact, there were no more than 10 per year until 2006, when there were 11 citations

      



from developing world patents to developed world patents. Patent citations from developing to developed world patents reached 201 in 2010 before reaching a peak of 1,256 in 2013, after which they subsided. From 2010 onwards, China accounted for the vast majority of these patent citations. Patent citations from developed to developing world patents finally exceeded 10 in 2008, then rose substantially, jumping more than an order of magnitude to 294 in 2017. Patent citations from US patents to developing world patents accounted for a large majority of this number during every year from 2007 to 2018, except in 2011. Figure 7.12 shows the aggregate importance of developed country wind power patents citing developing country patents, and vice versa. There are several gaps in this data. Importances of developed world patents cited by developing world patents are only continuous from 1990 onwards, and US patents cited by developing world patents from 2004. Importances of developing world patents cited by developed world patents are only continuous beginning in 2006, and Chinese patents cited by developed world patents later, from 2010. Nevertheless, several trends are noticeable. For developed world wind power patents cited by those in the developing world, importances vary over a range spanning almost two orders of magnitude, from a low of 0.59 in 1996 to a high of 53.81 in 2012. Over

Figure 7.12 Importance of citing wind power patents between developed and developing countries. Source: Authors’ analysis.



     .         .

this same time period, from 1996 to 2012, the dominant trend is upwards, though this rise is interrupted by several large reversals in direction. After 2012, the trend is downwards. For all spans of time during which developed world and US data are both available, cited US wind power patents account for almost all importance in the developed world. The pattern for cited developing world wind power patents is quite different. After 2006, when continuous cited developing world data begins, there is a rapid rise in importance to 2007, a substantial reversal in trend until 2009, and then a rapid rise to a maximum of 346.61 in 2012. Continuous data for cited Chinese wind power patents begins in 2010, accounting for almost all developing world patent importance until 2018. Figure 7.13 illustrates the aggregate importance of developed country wind power patents cited by developing country patents, and vice versa. The aggregate importance values of cited wind power patents are approximately an order of magnitude greater than those for citing wind power patents. There is continuous importance data for developed country wind power patents cited by those in developed countries. Aggregate importances range from a low of 0.25 to a peak value of 273.44, representing a difference of three orders of magnitude. From 1985 until 2009, these importances fluctuated frequently within a range from 0.25 to 8.87. Then, after 2009, they exhibited a rapid rise to the peak value of 273.44 in

Figure 7.13 Importance of cited wind power patents between developed and developing countries Source: Authors’ analysis.

      



2017 before falling in 2018. Data for the importances of cited US patents – incomplete until 2004 – showed a similar trajectory to that for the developed world, though US importances tended to be substantially lower. Aggregate importance data for developing world patents cited by those in the developed world are scarce prior to 2006, after which patent importances rose and fell substantially until 2009. From an aggregate importance of 74.06 in 2009, they then rose sharply, reaching a maximum of 9,092.57 in 2012, an almost 123-fold increase in only three years. After 2012, aggregate importance fell rapidly. The trend for importance of cited Chinese patents closely tracked that for cited developing world patents, though importances for these Chinese patents were consistently a bit lower. Figure 7.14 shows the mean importance of developed country patents citing developing country patents, and vice versa. Here, there is continuous importance data for developed country wind power patents citing developing country patents. Mean importance fluctuates from lows of zero in 1986 and 1989 to a high of 7.08 in 1994, but there is no clear directional trend. Figure 7.15 reveals the mean importance of developed country wind power patents cited by developing country patents, and vice versa. The

Figure 7.14 Mean importance of citing wind technology patents between developed and developing countries Source: Authors’ analysis.



     .         .

Figure 7.15 Mean importance of cited wind technology patents between developed and developing countries Source: Authors’ analysis.

only continuous mean patent importance data from 1985 to 2018 pertains to developed world patents cited by developing world patents. These mean values fluctuated between a low of 0.23 in 2004 and a high of 2.08 in 2009, but without any overall rising or falling trend. Values for US mean patent importances – for which data is discontinuous until 2004 – tracked developed world values relatively closely throughout the time period. Continuous mean importance data for cited developing world patents begins in 2006. These mean importances were consistently far higher than those for cited developed world patents, with maximum and minimum values of 9.45 and 3.00 respectively. From 2006 to 2009, these values dropped precipitously, recovering suddenly in 2010, then declining substantially until 2017, and finally recovering modestly in 2018. From 2010 onwards – when continuous Chinese data are available – mean importances for cited Chinese wind power patents closely tracked those for developing world patents, diverging substantially only in 2018.

Pharmaceutical Technology Figure 7.16 depicts the growth over time of pharmaceutical patents. Numbers of patents in the pharmaceutical technology cluster have grown

      



Figure 7.16 Trend in pharmaceutical patents over time Source: Authors’ analysis.

every year since 1970 with the exception of a brief pause from 1976 to 1979. Rates of growth accelerated noticeably after 1996, a year with 57,017 patents, to 138,934 by 2003, for a seven-year increase of almost 2.5 times. Then, from 2003 to 2015, they almost doubled again, reaching 263,059. In contrast to wind power patents, the decay in numbers of pharmaceutical patents did not begin until after 2015. These too are difficult to interpret, likely reflecting not a decline in pharmaceutical patents, but incomplete data reporting by national patent offices. Figure 7.17 shows the interjurisdictional patent citation rate between developed and developing countries for pharmaceutical technology. Although the number of patents in this technology field is a small subset of all patents across all technologies, as can be seen in Figure 7.16, it includes far more patents than does wind power (Figure 7.10). The pattern in Figure 7.17 is strikingly similar to that for all technological fields (Figure 7.2), replicating the overall rise-fall-rise pattern. However, as for wind power patents, the rate of interjurisdictional citation of pharmaceutical patents has shifted higher during every time period except from 1975 to 1978. Figure 7.18 illustrates trends in the aggregate citations between pharmaceutical patents originating in the developed and developing world. The number of raw patent citations in pharmaceutical technology made by patents in the developed world to those in the developing world rose steadily, roughly quadrupling during the two decades from 1985 to



Figure 7.17

     .         .

Rates of interjurisdictional citation for pharmaceutical technology

Source: Authors’ analysis.

Figure 7.18 countries

Pharmaceutical patent citations between developed and developing

Source: Authors’ analysis.

2005. Then, the growth trajectory accelerated, rising by more than an order of magnitude, from 1656 in 2005 to 22,935 in 2017. Citations of US pharmaceutical patents by developing world patents rose in a similar pattern, consistently accounting for about half of all developed world citations. With the exception of 1986, complete citation data for developing world pharmaceutical patents cited by developed world patents exists for

      



Figure 7.19 Importance of citing pharmaceutical patents between developed and developing countries Source: Authors’ analysis.

the entire time period from 1985 to 2018. This trend shows far more volatility than that for cited developed world patents. Until 1995, it lags behind cited developed world patents, pulling even in 1995, but then falling behind again – sometimes far behind – until pulling even for a second time in 2006. From 2004 until 2007, these citations grew more than 35-fold, from only 215 to 7,660, they fell back down to 5,206 in 2009, before reaching a peak of 58,849 in 2014, and then decreased substantially thereafter. Data for cited Chinese pharmaceutical patents is available from 2005 onwards. Starting with a paltry two citations in 2005, these Chinese citations rose very rapidly, accounting for a large majority of developing world citations from 2011 to 2017. Figure 7.19 shows the aggregate importance of developed country pharmaceutical patents citing developing country patents, and vice versa. Despite frequent fluctuations, the aggregate importance of developed world pharmaceutical patents that cite developing world patents generally rose from 1985 to 1997, reaching a peak of 5,249.41, subsided thereafter down to 1,398.12 in 2007, reversed direction again after 2007 to climb back to 2,910.64 in 2012, and then declined steadily until 2018. The aggregate importance of US patents closely matched this pattern, and always accounted for a large majority of developing world aggregate importance.



     .         .

Figure 7.20 Importance of cited pharmaceutical patents between developed and developing countries Source: Authors’ analysis.

Including small fluctuations in value, the aggregate importance of developing world pharmaceutical patents that cite developed world patents remained very low in the two decades from 1985 to 2005. Starting in 2006, these importance values rose substantially. From 0 in 2005, importance rose to 644.22 in 2007, declined to 293.79 by 2009, and then rocketed upwards, exceeding the value for developed country patents citing developing country patents by 2011, and reaching a maximum of 9,396.83 in 2012, before declining sharply again. Aggregate Chinese importance remained very low until 2009, and then closely tracked the value for all developing country patents, accounting for almost all their importance. Figure 7.20 illustrates the aggregate importance of developed country pharmaceutical patents cited by developing country patents, and vice versa. The aggregate importance values of cited pharmaceutical patents are approximately an order of magnitude greater than those for citing pharmaceutical patents. The trend in aggregate importance of developed country pharmaceutical patents cited by developing country patents was steadily upwards from 382.38 in 1985 to 25,645.84 in 2017, a prodigious rise of almost two orders of magnitude. The aggregate importance of US patents rose in a similar trajectory. For most of the period from 1985 to 2018, the US value accounted for a majority of developing world aggregate patent importance, but its fraction of importance did slip below 50 per cent from 2005 to 2014.

      



The trend for developing world pharmaceutical patents cited by developed world patents also exhibited a general rising trend, but experienced more fluctuations. Although slightly above the aggregate importance for developed world pharmaceutical patents in 1987, the importance of developing world patents slipped below the developed world level from 1988 until 1991. Beginning in 1992, it rose well above the aggregate importance of developed country patents, declining again after 1995 and drawing close to even with developed country importance in 1997 and 1998. After this, aggregate importance of cited developing world patents remained consistently above the developed world level, and usually at about an order of magnitude above. From a very low aggregate importance of 11.71 in 2005, it reached a high of 186,952.43 in 2014, for an increase of more than four orders of magnitude in less than a decade. From 2011 to 2016, aggregate Chinese importance made up a majority of developing world importance. Figure 7.21 shows the fvmean importance of developed country patents citing developing country patents, and vice versa. After an initial decline, the mean importance of developed world pharmaceutical patents that cite those from the developing world fluctuated above and below 4.00 from 1989 to 1997, after which it declined steadily from 5.14 in 1997 towards zero in 2018. The mean importance of citing US pharmaceutical patents was consistently higher than that for developed country patents. It also declined after 1985, losing almost half its importance by 1998, then rebounded to reach a new high in 1995, declined sharply in 1996, and then reached another new high in 1997. With several fluctuations, the trend in mean importance from 1997 onwards declined steadily towards zero in 2018. Mean developing world patent importances stayed near zero from 1985 until 2005. Next, they spiked upwards to 0.10 in 2006, declined modestly until 2009, and then rose sharply again, reaching 0.28 in 2011, before declining towards zero in 2018. The pattern for Chinese patents citing developed world patents started higher than that for developing world patents, at 0.43 in 2006, converging on the developing world patent importance trend by 2012. Figure 7.22 reveals the mean importance of developed country pharmaceutical patents cited by developing country patents, and vice versa. The trend in mean importance for developed world pharmaceutical patents cited by developing world patents remained quite flat,



     .         .

Figure 7.21 Mean importance of citing pharmaceutical patents between developed and developing countries Source: Authors’ analysis.

Figure 7.22 Mean importance of cited pharmaceutical patents between developed and developing countries Source: Authors’ analysis.

almost always just above 1.0, from 1985 to 2018, with the exception of a modest peak in 1997. The trend for US patents was quite similar, except for a modest decline from 2006 to 2011. With the exception of 1985, the mean importances for developing country patents cited by developed world patents were consistently about five times to an order of magnitude larger than those for developed world patents. Punctuated by substantial fluctuations, developing country patent

      



importances, though fluctuating substantially, generally rose rapidly after 1985, reaching a maximum of 22.35 in 2000, and then generally declined thereafter. The mean importance of Chinese pharmaceutical patents started well below that for developing world patents, then converged on the developing world values, before falling well below them from 2016 onwards.

Worldwide Knowledge Flow among All Countries One way to analyse international patent knowledge flow is to use a radial diagram in which almost all countries and jurisdictions in the world are arrayed around the circumference of a circle.50 The radial diagram in Figure 7.23 shows the flow of citations to and from developed and developing countries measured by importance. All citations from patents of one country to those of another country are summed together. The thicker the lines between countries, the more citations between them, relative to other links. In the radial diagrams shown in this chapter, we show 75 per cent of all the links measured by importance. We removed a portion of the links to highlight the key patterns. We adopt this practice because including all patent citation links in a radial diagram can make major patterns of knowledge flow difficult to see due their high density. The dark grey two-letter codes represent developed countries. The light grey colours represent developing countries. The two-letter code ‘WO’ represents the PCT, which accepts patent applications from applicants in almost all countries; because it belongs neither to the developed nor the developing world, the WO is situated in Figure 7.23 between the two categories of countries or regional filing authorities. Because the PCT is an international filing system, not a final destination of patent applications, it is difficult to construe as a node of knowledge flow as such; with that caveat, it is included for completeness’ sake. Figure 7.24 illustrates patent knowledge flow from two perspectives: from developed countries’ patents to developing countries’ patents (the left-hand radial diagram); and, from developing countries’ patents to developed countries’ patents (the right-hand radial diagram). The figure’s lines do not represent the sum of links as they do in Figure 7.23. 50

See http://well-formed.eigenfactor.org/radial.html.



Figure 7.23

     .         .

Radial diagram of international knowledge flow

Source: Authors’ analysis.

Nor do they include citations from developed world patents to other developed world patents or developing world patents to other developing world patents. The presence or absence of lines linking countries does represent whether or not there is substantial important patent knowledge flow. The radial diagram in Figure 7.24 reveals several interesting patterns. Not all developed countries possess a substantial number of patents that are cited regularly by those in developing countries, as discussed earlier in the chapter. Less surprisingly, the converse is also true; not all developing countries possess significant holdings of patents. Patent knowledge flows from the patents of 20 developed countries to the

       



Figure 7.24 Directional knowledge flow between developed and developing world countries Source: Authors’ analysis.

patents of 35 developing countries. By contrast, patent knowledge from the patents of only 21 developing countries flows to the patents of only 16 developed countries. In other words, not only is there less patent knowledge flow from developing countries’ patents to developed countries’ patents, there is substantially less participation by patent systems in both developed and developing countries in knowledge flow from developing to developed countries. One stark contrast is that a much higher fraction of developed world countries’ patent systems participate in patent knowledge flow in both directions (that is, developed to developing world and developing to developed world) than do those in the developing world. This reveals in a new light considerable opportunties for growth in



     .         .

Figure 7.24 (cont.)

contributions developing world countries may make to worldwide patent knowledge flow – contributions capable of increasing for all technological innovation and the benefits therefrom.

Discussion Knowledge Flow between More and Less Developed Countries Fluctuating Interjurisdictional Citation Rates Interjurisdictional patent citation generally grew robustly between 1970 and 2000, roughly quadrupling from about 10 per cent in 1970 to

      



almost 40 per cent in 2000 (Figure 7.2). This parallels the growth in patents during the same period (Figure 7.1), which also approximately quadrupled. It is also consistent with the increasing availability of scientific, technical and patent documents, data, and other information with improvements in information technology, such as personal computers, the Internet, and the World Wide Web. However, the similarity between interjurisdictional citations and patent numbers ends around 2000, after which patent numbers not only continued to increase, but accelerated, reaching almost 6,000,000 in 2017, while citations decreased sharply, halving by 2005. Interjurisdictional citations did subsequently increase again, including a particularly rapid rise from 2010 to 2012, before remaining stable at about 30 per cent. The sharp decline and recovery in interjurisdictional citations defies easy explanation. The 2000–2003 global recession might seem a likely candidate to have influenced the drop in citations after 2000. However, if so, it is odd that rates of patent citations declined while numbers of patents continued to climb. The drop in citations did not correspond with the WTO–TRIPS amendments, which were finalized in 1995, nor with any other major changes in international, US, or European patent law. Similarly, it is difficult to explain the subsequent recovery in the rate of patent citations, which also did not coincide with any obvious explanatory external legal or economic event. Another possible explanation is missing patent data, but there is no evidence of this. It is possible that the post-2000 decline in interjurisdictional patent citation rates is a consequence of the natural accumulation of citations over time. This is certainly true for numbers of patent citations. As a patent ages, it tends to attract more citations. Thus, younger patents tend to have attracted fewer citations than older ones. However, it is not obvious why the rate of interjurisdictional patent citation rates would be similarly affected. Perhaps it is more likely that a patent will be cited by other patents in the same jurisdiction due to their proximity to one another. But, if this were so, then the rapid recovery in the rate of interjurisdictional patent citations after 2010 is confounding.

Towards More Balanced Knowledge Flow Until 2005, there were considerably more developed world patent citations to developing world patents than vice versa, though overall numbers of worldwide patent citations were fairly stable at a much lower level than in recent years (Figure 7.3). After 2005, two striking changes



     .         .

occurred. First, the total number of patent citations worldwide rocketed upwards by about two orders of magnitude. Second, the pattern of patent citation between developed and developing countries reversed, with developing country patents citing developed country patents much more frequently than vice versa. In fact, this pattern reversal occurred despite the rapid growth in patent citations of developing countries by developed countries. After 2009, China accounted for most of the increasing citations from developing world patents. One implication of these patterns is that, in recent years, patent systems in the developing world have become much more aware of the relevance to their own patents of developed world patents. Another is that the same phenomenon applies in the developed to developing world direction, but to a lesser extent. Nevertheless, the citation data from 2015 onwards suggest that developing world patents may be reducing their citations of those from the developed world while the latter continues to increase its citations of developing world patents. If these trends were to continue, the magnitude of citations between developed and developing world patents would appear likely to equalize, at least briefly. Developing countries appear to be increasing rapidly as sources of technical information for the developed world, and to be relying less themselves on technical information from the developed world. This suggests that, while technical knowledge has been flowing robustly from developed countries to developing countries in recent years, this dependency has begun to lessen, leading towards more balanced knowledge flow.

A Patent Importance Perspective A Cautionary Illustration It may be preferable to rely on importance measures than on raw citations, as the latter can yield misleading estimates of technical knowledge flow. Consider the citation patterns between Chinese and US patent documents. Chinese patent documents have cited US patents or patent applications on 2,209,396 occasions, while US patent documents have cited Chinese documents only 592,154 times. This represents a major imbalance because Chinese patent documents cite US patent documents 3.73 times more often than US patent documents cite Chinese patent documents. However, when the importance of these citations is taken into account, this skewed citation pattern becomes more uneven, reaching a difference of almost an order of magnitude because the importance of

      



Chinese-to-US citations is 15,261,498 while the importance of US-toChinese citations amounts to only 861,941. By this measurement, the importance of the US patents cited by Chinese patents is 17.7 times as great as the importance of the Chinese patents cited by US patents. One interpretation of these results is that US technical knowledge embodied within patents has flowed much more robustly to, rather than from, China. Furthermore, in selecting accurate measurements of this knowledge flow, those relying on raw citations may vastly underestimate the magnitude of this skewed knowledge flow. It may be advisable to rely more on importance-weighted citations than raw citations for measuring knowledge flow.

Estimating Knowledge Flow from Importance Aggregate importance of citations adds a different perspective. We examined aggregate importance from two perspectives: developed world patents that cite developing world patents (Figure 7.4) and vice versa (Figure 7.5). In the first case, developed world patents maintained far higher importance from 1985 to 2010, after which the importance of developing world patents raced ahead. The sustained rise in importance of developing world patents citing developed world patents occurred after 2005. Throughout the time period measured, US patents accounted for the vast majority of developed world patent importance. Similarly, after 2009, Chinese patents contributed the vast majority of developing world patent importance. One implication of this pattern is that citation of developed world patents has recently become a very strong indicator of importance for developing world patents. In fact, since 2010, citation of developed world patents has tended to be strongly associated with the importance of developing world patents. By contrast, the relatively modest fluctuations in the importance of developed world patents that cite developing world patents suggests little effect on the importance of the former. These results also reveal the disproportionate importance of US and Chinese patents within the wider categories of developed and developing world patents, respectively. Whether developed world patents were cited by developing world patents seems not to have had a substantial effect on the importance of the former. There may have been an increase in the most recent years, but even this rise has been modest. Again, it is the importance of developing world patents that appears to have been affected; beginning



     .         .

after 2000, and especially after 2009, developing country patents cited by those in developed countries have increased prodigiously in importance. Again, US patents have consistently accounted for much of developed world patent importance, and, after 2009, Chinese patents have played a similar role among developing world patents. As with citations flowing in the other direction, being cited by developed world patents appears to have been a strong indicator of developing world patent importance, whereas the importance of developed world patents seems to have been affected little by citations from developing world patents. An examination of mean patent importances tells a different story. Mean importances of developed world patents citing those in the developing world dwarfed those of developing world patents citing developed world patents from 1985 until 2010, though this disparity began to decrease steadily after 1997, and mean importances approached convergence from 2010 onwards (Figure 7.6). Mean importances for US patents were substantially higher than those for all developed world patents, but showed a very similar trend to the latter. Similarly, mean Chinese patent importances were above those for developing countries, but held to a similar trajectory. There was a reversal of roles for developing country patents cited by developed world patents. After 1986, the mean importance of developing world patents remained well above the mean importance of developed world patents cited by developing world patents (Figure 7.7). The mean importance of US and developed world patents was very similar throughout the time period measured, whereas mean importances of Chinese patents were initially much higher than those for the developing world, but rapidly converged with them. The results for mean patent importance provide little evidence that developing world patents that cite developed world patents tend to be more important, though the modest bump in importance after 2009 might suggest an incipient change. On the other hand, mean importances of developing world patents cited by developed world patents have been noticeably higher than vice versa.

Climate Mitigation Technology Knowledge Flow In addition to general trends of citation and importance spanning all areas of technology, we also examine two particular technological fields: wind power and pharmaceuticals. First, we explore wind power

      



technology. We test the hypothesis that the trajectory of climate-change mitigation knowledge flow increased (especially flow from developed to developing countries) after the 1997 Kyoto Protocol51 and the 2015 Paris Agreement52 on climate change.

Evidence of Knowledge Flow in Wind Power Technology The relatively small number of patents in the field of wind power technology makes it challenging to make fine-grained comparisons to overall trends in patent citations and importance. More generalized comparisons must suffice, and, at this level, the trends and patterns for wind technology patents generally resemble those for all technologies combined. The growth trajectory is similar and strongly upwards in both wind power patents (Figure 7.9) and all patents (Figure 7.1), although the rapid acceleration of wind patents lags several years behind that observed for both all patents and pharmaceutical patents (Figure 7.16). The pattern in rates of interjurisdictional citation (Figure 7.10) is similar for both, though the upward shift in rates for wind power patents suggests more knowledge flow between the developed and developing world in this technological field, and a generally upward trend after 2010. Patterns of citations (Figure 7.11), aggregate importance (Figures 7.12 and 7.13), and mean importance (Figures 7.14 and 7.15) are not meaningfully different from those for all patents. And, while there is no clear evidence that the Kyoto Protocol influenced any of the metrics used to characterize wind technology patent knowledge flow, Figure 7.10 does illustrate a sharp reversal in the trajectory of interjurisdictional citation rate for wind power patents one year after the successful 2015 negotiation of the Paris Agreement. This latter observation must be treated with caution, however, because it is the product of very recent patent data that is likely incomplete, and, thus, may constitute a biased sample. Furthermore, correlation is not causation, so one must be circumspect about attributing this increase to the Paris Agreement. The paucity of recent data will be solved by the passage of time, which will allow more complete data, accumulation of citations and more accurate estimates of importance. 51

52

The Kyoto Protocol is a 1997 treaty intended to implement the 1992 United National Framework Convention on Climate Change. (See https://unfccc.int/process-and-meet ings/the-kyoto-protocol/what-is-the-kyoto-protocol/what-is-the-kyoto-protocol.) The Paris Convention is a 2015 treaty intended to implement the 1992 United National Framework Convention on Climate Change. (See https://unfccc.int/process-and-meet ings/the-paris-agreement/the-paris-agreement.)



     .         .

Pharmaceutical Technology Knowledge Flow Defining ‘Pharmaceutical’ The WTO–TRIPS Agreement entered into force on 1 January 1995.53 In addition to establishing a minimum level of patent protection in member countries, it represented a step towards the standardization of patent law worldwide. One technology field particularly affected by this standardization was pharmaceutical drugs. Many in the pharmaceutical industry hoped that the TRIPS Agreement would increase the potency of patent protection around the globe, while many, including those in developing countries, wished for enhanced access to powerful medicines and information about how to make and use them. It is challenging to quantify the latter knowledge flow of technical information. As described in the above section on methodology,54 we use eigenvector centrality to attempt to measure knowledge flow among countries, including in the field of pharmaceutical drugs. A challenge in doing so successfully relates to the complexity of ‘pharmaceutical drugs’ as a category of technology. Our method of classifying technologies relies on the hierarchical structure of patent citation network our method produces. In this network, there are approximately 100 million relatively discrete technology clusters. These clusters tend to be nested within larger, more inclusive, technology clusters, with all clusters ultimately nested, either directly or indirectly, within the largest, most inclusive, cluster, which is the entire network itself. These clusters may be thought of as ‘natural’ technology groups because they form as a result of the structure of the network. This contrasts with technology groups resulting from classification systems such as the IPC and CPC schemes, which are human constructs into which each patent application is placed, often as the consequence of a human decision. The clustering method of technology classification we employ is not congruent with schemes such as the IPC or CPC. Rather, traditional categories of technology are often dispersed among IPC or CPC categories. In other words, our classification of ‘pharmaceutical drug’ is unique. Consequently, we focused our analysis on a particular natural technology cluster, not any derived from the IPC or CPC schemes. In our network 53

54

The WTO TRIPS Agreement allowed staged implementation by different groups of WTO members. Thus, many developing countries did not adopt all TRIPS provisions until years after its initial entry into force. P. 224 supra.

      



we have defined a large natural cluster that comprises the vast majority of patents traditionally classified as falling within pharmaceutical technology.55 We use this cluster to represent ‘pharmaceutical technology’ in the following analyses.

Evidence of Knowledge Flow in Pharmaceutical Technology The World Health Organization (‘WHO’) has stated that, of all the WTO agreements, TRIPS ‘is expected to have the greatest impact on the pharmaceutical sector and access to medicines’.56 Although more than 40 countries lacked patent protection for pharmaceuticals prior to the TRIPS negotiations,57 a result of the Agreement was to make availability of such protection mandatory, though the necessity to comply was phased in for certain countries and TRIPS does afford WTO members some degree of discretion in implementation,58 (and least developed country members have no such obligations until at least 2033). The TRIPS Agreement did come into force in 1995, but there was no single date upon which all pharmaceutical inventions became patent eligible subject matter. This complicates the interpretation of citation and PVweighted citation because some developing country members brought in patent protection for pharmaceuticals in 1995, or even beforehand, while others waited until as late as 2005 to do so. Our analyses do offer some evidence consistent with the suggestion that knowledge flow in pharmaceutical technology increased after relevant TRIPS obligations began to apply from 1995. The increase in pharmaceutical patents accelerated soon after 1995, maintaining a much higher growth rate until 2015 (Figure 7.16). However, the trend for all patents across all technological fields showed an even more impressive growth rate after 1996 (Figure 7.2), suggesting that pharmaceutical technology, despite its acceleration, has actually been a laggard compared with the more general trend after 1995. Although the rate of interjurisdictional citation for pharmaceutical technology (Figure 7.17) has closely mirrored the shape for all patents (Figure 7.3), from 1979 onwards the rate of cross-border citation in pharmaceutical patents has been substantially higher. This suggests that knowledge flow is especially strong in pharmaceutical technology, and, if the post-2005 trend is 55 56 57 58

This cluster roughly corresponds to CPC A61K. See www.who.int/medicines/areas/policy/wto_trips/en/. See www.who.int/medicines/areas/policy/wto_trips/en/. See www.who.int/medicines/areas/policy/wto_trips/en/.



     .         .

indicative, likely to grow stronger. On the other hand, pharmaceutical patent trends in citations (Figure 7.18), aggregate importance (Figures 7.19 and 7.20) and mean importance (Figures 7.21 and 7.22) do not appear to diverge meaningfully from those for patents across all technological fields. It should be noted that any causal effects the TRIPS Agreement might have had could have been masked by the staggered entry into legal force in various countries of its full legal obligations germane to pharmaceutical drugs.59

Conclusions Employing eigenvector centrality and hierarchical network methods to examine technical knowledge flow offers advantages over traditional methods based only on raw patent citations. A primary advantage is that neither patent numbers nor raw citations are capable of reflecting the importance of the technical knowledge that does flow from patent to patent, party to party, and country to country. Developed country patents supply a disproportionate amount of patent citations compared to patents from developing countries, and the importance of technical knowledge thereby transferred is disproportionately high. However, we observe some evidence that this imbalance in knowledge flow may be beginning to reverse itself, as developing country patents have become increasingly cited by those in developed countries, and the importance of developing world patents has grown. If this observation does foreshadow a sustained shift in patent citations and patent importance, technical knowledge via patents from the developing to the developed world may soon become an important factor in global knowledge flow. Our results failed to discern evidence that the Kyoto Protocol has spurred knowledge flow in climate-mitigation technology, but they did generate some preliminary evidence consistent with an effect by the Paris Agreement. We did find evidence consistent with the possibility that the TRIPS Agreement may have spurred pharmaceutical (and more general) knowledge flow. One striking observation apparent from the analyses in this chapter is that US patents make up a huge proportion of the patent citations, patent importance and technical knowledge flow within the developed world. 59

See www.wto.org/english/tratop_e/trips_e/factsheet_pharm04_e.htm.

       



Similarly, China has recently come to dominate patent citations, patent importance and technical knowledge flow within the developing world. These two phenomena may change over time, leading to less influence by one or both countries. In the meantime, it may be possible to use the US and China as relatively accurate proxies for patent analyses involving the developed and developing worlds, respectively, without sacrificing inordinate resolving power. The use of worldwide patent data and hierarchical mapping network approaches offer new, and potentially useful, insights into the flow of technical knowledge among countries and other patent jurisdictions, within specific technologies, and over defined periods of time.

Bibliography Albert, Michael B., Avery, D., Narin, Francis and McAllister, Paul R., ‘Direct Validation of Citation Counts as Indicators of Industrially Important Patents’ (1991), Res. Pol’y 20: 251. Allison, John R., ‘Valuable Patents’ (2004), Georgetown Law Journal 92, 435: 451. Carpenter, Mark P., Francis Narin and Woolf, Patricia, ‘Citation Rates to Technologically Important Patents’ (1991), 3 World Patent Information 3: 160. Hall, Bronwyn H., Jaffe, Adam B. and Trajtenberg, Manuel, ‘Market Value and Patent Citations: A First Look’, 2000, National Bureau of Economic Research, Working Paper No. 7741. Hall, Bronwyn H., Jaffe, Adam and Trajtenberg, Manuel, ‘Market Value and Patent Citations’ (2005), Rand Journal of Economics, 16: 36. Harhoff, Dietmar, Narin, Francis, Scherer, F. M. and Vopel, Karin, ‘Citation Frequency and the Value of Patented Inventions’ (1999), Econ. & Stat 81, 511. Lanjouw, Jean O. and Schankerman, Mark, ‘Characteristics of Patent Litigation: A Window on Competition’ (2001), Rand Journal of Economics 32, 1: 129–151. Lanjouw, Jean O. and Schankerman, Mark, ‘Enforcing Patent Rights: An Empirical Study’, in Empirical Economics of Innovation and Patenting, 2003, p. 14, available at LES Research Online, http://eprints.lse.ac.uk/58574/. Narin, Francis, Noma, Elliot and Perry, Ross, ‘Patents as Indicators of Corporate Technological Strength’ (1987), Res. Pol’y 16: 143. Page, Lawrence, Brin, Sergey, Motwani, Rajeev and Winograd, Terry, ‘The PageRank Citation Ranking: Bringing Order to the Web, Stanford InfoLab’, 1999, available at http://ilpubs.stanford.edu:8090/422/1/1999-66.pdf. Palangkaraya, Alfons, Webster, Elizabeth and Jensen, Paul H., ‘Misclassification Between Patent Offices: Evidence From a Matched Sample of Patent Applications’ (2011), Review of Economics & Statistics 93: 1063.



     .         .

Sampat, Bhaven N. and Ziedonis, Arvids A., ‘Patent Citations and the Economic Value of Patents’, in Handbook of Quantitative Science and Technology Research 277, H. F. Moed, W. Glänzel, and U. Schmoch (eds), Berlin: Kluwer Academic Publishers, 2004. Seres, Stephen, ‘Analysis of Technology Transfer in CDM Projects, 7’, 2008, available at http://cdm.unfccc.int/Reference/Reports/TTreport/TTrep08.pdf. Torrance, Andrew W. and West, Jevin D., ‘All Patents Great and Small: A Big Data Network Approach to Valuation’ (2017), Virginia Journal of Law & Technology 20, 3: 466–504. Trajtenberg, Manuel, ‘A Penny for Your Quotes: Patent Citation and the Value of Innovations’ (1990), Rand Journal of Economics 21: 172–187. West, Jevin D. and Vilhena, Daril A., ‘A Network Approach to Scholarly Evaluation’, in Bibliometrics & Beyond: Metrics-Based Evaluation, Blaise Cronin and Cassidy R. Sugimoto (eds), Cambridge, MA: MIT Press, 2014, pp. 151–166. West, Jevin D., Bergstrom, Theodore C. and Bergstrom, Carl T., ‘The Eigenfactor Metrics: A Network Approach to Assessing Scholarly Journals’ (2010), College & Research Libraries 71: 236. West, Jevin D., Wesley Smith, Ian and Bergstrom, Carl T., ‘A Recommendation System Based on Hierarchical Clustering of an Article-Level Citation Network’ (2016), IEEE Transactions on Big Data 2, 2: 113–123. Zhong, Fa Ma, ‘The Effectiveness of the Kyoto Protocol and the Legal Institution for International Technology Transfer’ (2010), Journal of Technology Transfer 37: 75, 83.

8 Sources of Knowledge Flow between Developed and Developing Countries                 ,                                  Abstract This chapter provides a long-term view on the sources of knowledge flow between developed and developing nations. It relies on patent data to explore three potential sources: R&D collaboration, technology sourcing, and technology transfer. All three sources provide a very consistent message. First, knowledge flows with East Asia, particularly China, are occurring more frequently. Second, knowledge flows are increasingly concentred in information and communication technologies. Third, the United States & Canada had traditionally larger patenting activity with Asia than Europe, but the share of activity between Europe and Asia has been increasing in recent years. Fourth, larger patenting activity between the United States & Canada and Asia implies that the US–Canada region is more likely to benefit from reverse knowledge flows as China progresses towards becoming a technological leader.

Introduction* R&D globalization is a defining feature of modern innovation systems and a major driver of productivity growth (Eaton and Kortum, 1996; Thomson and de Rassenfosse, 2016). Innovating firms set up R&D labs abroad or cooperate with foreign partners in order to tap into an ever more globally dispersed knowledge frontier and to reduce R&D costs (Harhoff et al., 2014). From the viewpoint of host countries, policymakers are keen to attract foreign R&D. The knowledge that flows from foreign R&D may be * Part of the work outlined in this chapter was funded thanks to a grant by the Swiss National Science Foundation (application number 100018_169584, entitled ‘Globalization of R&D: Technology Cluster, Performance and Risk’). The authors are grateful to Andy Toole and Jayashree Watal for helpful comments.





          .

appropriated by local firms and raise their productivity level. If appropriation takes place without local firms contributing to the cost of the research, one talks of ‘knowledge spillovers’. Appropriation can also occur in the context of a transaction on the market for technology (Arora et al., 2004). The existence of cross-border knowledge flows has been well documented in the academic literature. Most of the literature tracks knowledge flows by means of patent citations, where citations are measured between a citing and a cited country, industry or firm (see, e.g., Jaffe and Trajtenberg, 1999; Maruseth and Verspagen, 2002; Peri, 2005). The vast majority of studies focus on developed economies – literature on developing economies is scarcer (see, e.g. Svensson, 2007). For developing countries, attracting foreign knowledge from more developed regions is both crucial and challenging. It is crucial because knowledge flows from the technology frontier are a key source of learning (Griffith et al., 2004; Seliger, 2016). It is challenging because developing countries lack the knowledge base and infrastructure to fully benefit from them. This chapter contributes to the literature on cross-border knowledge flows in two ways. First, it puts the spotlight on flows between developed and developing economies. Second, it focuses on some potential sources of knowledge flows rather than on documenting their existence. We explore three potential sources: R&D collaboration, technology sourcing and technology transfer.1 We quantify these sources by means of patent data. In contrast to prior studies that focus on patents filed at selected patent offices in the United States or Europe, our analysis considers 52 patent offices around the world. R&D collaboration is measured by looking at patents that are coinvented or co-applied by parties in two different countries. Technology sourcing is measured by patents that are applied by a party in one country but invented in another country. Finally, knowledge transfer is measured using information on the transfers of ownership (i.e. sales) of patents. We find that there has been a large shift towards R&D collaboration and technology sourcing with Asian countries (especially China) in the last two decades. Asia is today the only significant area for R&D collaboration and technology sourcing for Western economies. We also find that R&D collaboration and technology sourcing mainly take place in the 1

Other sources of knowledge transfer exist, such as the mobility of R&D personnel (e.g. Lenzi, 2010). Our analysis is silent on this and other sources.

                        



information and communication technology (ICT) sector and especially in computer technology. Regarding patent transfers, we document an important increase in the last decade, although overall levels remain low. East Asia is the largest recipient region among developing countries. Patent transfers from developing to developed economies also exist, although the volumes are more modest, reaching about a third of the volumes of the opposite direction.

Method and Data In contrast to studies that seek to assess the effect of knowledge flows on the sending or receiving country, this chapter focuses on the sources of knowledge flows. It is well-documented that interpersonal and interfirm collaboration raise knowledge flows between the parties involved (Agrawal et al., 2006; Gomes-Casseres et al., 2006; Singh, 2005). We discuss the three potential sources in turn.

Measuring R&D Collaboration with Patent Data Cross-border R&D collaboration occurs when firms from different countries work together on an invention or when inventors from different countries work together. These aspects can be measured by looking at applicant–applicant and inventor–inventor pairs of the patent documents, respectively (see Guellec and de la Potterie, 2001; Picci, 2010; Picci and Savorelli, 2018). In the context of the present study, one applicant (inventor) must come from a developed country and the other applicant (inventor) from a developing country. Consider for example the case of patent application number FR2965057A1 filed on 22 September 2010. This patent is co-applied by two organizations, namely the Soils and Fertilizers Research Institute (Vietnam) and the Institut de Recherche pour le Développement (France). This patent is most likely the result of R&D collaboration between these two institutes. Note that the term ‘collaboration’ is more appropriate for inventor– inventor pairs than for applicant–applicant pairs. The fact that a patent is co-applied does not necessarily imply that applicants are on equal footing. The patent could be fully developed by a subsidiary and coapplied by the subsidiary and the headquarter. Our data do not include ownership information and we are thus unable to assess whether patents arise from collaboration between two related entities. As we will see,



          .

however, insights obtained from applicant–applicant pairs are very similar to insights obtained from inventor–inventor pairs. Most research in the field relies on the count of patents filed at major patent offices, such as the European Patent Office (EPO) or the United States Patent and Trademark Office (USPTO). However, the patenting activity at these offices may not necessarily be representative of the overall situation across the globe. By contrast, the present analysis exploits the worldwide count of priority filings put forward by de Rassenfosse et al. (2013). This patent count identifies all priority patent applications filed worldwide and assigns them to the country of residence of the applicant and/or inventor. A priority patent application is the first patent application that is filed to protect an invention, anywhere in the world. The implementation of this count is quite challenging because of missing data on the inventors’ and applicants’ countries of residence. The authors have proposed a technical solution for recovering missing information on countries of residence that involves exploiting patent family information. Briefly, the technical solution involves looking for information on the countries of residence of applicants and inventors in the subsequent patent applications that belong to the same family.2 Subsequent filings are typically made in other jurisdictions (i.e. at other patent offices) than the priority filing but this need not be the case. The ‘worldwide count of priority filings’ considers 52 patent offices and seeks to impute potential missing country information (for more details, see A8.2).

Measuring Technology Sourcing with Patent Data We measure technology sourcing by considering patents for which applicants and inventors are located in different countries. In our context, one country needs to be a developed country and the other needs to be a developing country. Such patents can indicate R&D offshoring or external R&D contracting. In the case of R&D offshoring, the inventor works for a foreign subsidiary, whereas the patent is filed by the mother company. In the case of R&D contracting, a firm contracts out research to an R&D supplier. 2

A family of patents is a set of patents that are related to each other through common priority patent applications. The original data in de Rassenfosse et al. (2013) have recently been updated and improved thanks to an EPFL-ETH Zurich project funded by the Swiss National Science Foundation.

                        



Consider for example patent application number DE102011014441A1, filed on 18 March 2011. The sole applicant is the Germany-based automotive company Daimler AG and the inventors are all located in Pune, India. The Daimler group has an R&D centre in Pune, and the patent application is therefore the result of R&D offshoring. The literature usually identifies three main motives for technology sourcing. First, the technology-seeking motive relates to firms tapping into frontier knowledge that is not available in their home country. Second, the market-seeking motive relates to firms seeking to enter new markets and involves adapting home technology to local specificities. Third, the costreduction motive relates to performing R&D tasks in a country with cheaper labour (Chung and Alcácer, 2002; Dunning and Lundan, 2008). The data at hand do not allow us to distinguish between the various motives but we note that one or more of these motives may be at play. The count again relies on priority filings with imputed applicant and inventor country information. We exclude patents that were also coapplied (i.e. patent applications that have been filed by applicants from both the developing and developed countries). This restriction ensures that we exclude patents arising from technology collaboration and, therefore, that we identify more accurately cases of technology sourcing.

Measuring Technology Transfer with Patent Data We measure technology transfer using detailed information on the transfers of ownership, i.e. the sale of rights. We rely on the register of the European Patent Office (EPO) for patent applications, and on the register of the German Patent Office (DPMA) for granted European (EP) patents. Indeed, once a European patent application has been granted by the EPO, it is transformed into a bundle of national patents. This implies that, before grant, a transfer of patent ownership should be registered at the EPO, while after grant, the transfer has to be registered in national offices of countries in which the patent has been validated.3 Germany is the country with the largest proportion of validated EP patents (95 per cent in 2003). It is therefore the most complete source for post-grant changes of ownership of European patents (Harhoff et al., 2009). It should be mentioned that registering a transfer of patent ownership is mandatory neither at the EPO, nor at the DPMA. There are 3

It should be noted that it is still possible to register the transaction at the EPO during the nine-month period following the grant of the patent.



          .

nevertheless strong incentives for patent purchasers to register patent acquisitions, as the lack of registration may have severe consequences. Indeed, in many European countries, the non-registration of patent transfers considerably hinders the enforcement of patent rights in case of litigation (see Ciaramella et al., 2017 for a more detailed discussion). In comparison with the amounts at stake before courts, the costs of registering a transfer of ownership are negligible.4 Note that the countries of patent protection (in our setting, European countries), and of patent ownership are distinct. For example, it is possible for a European patent to be owned by a US firm and transferred to a Chinese firm. Thus, we will be able to provide a global view of patent transfers even using EP patents alone. Consider for example patent application number EP1905886A1, filed on 2 April 2008. This patent was developed by the Germany-based company AMTEC Wäschereimaschinen GmbH. It was then transferred on 7 July 2010 to Supreme Machinery Manufacturing Co. LTD, based in Thailand. Empirical large-scale research investigating the transfers of patent ownership is recent. It started with Serrano (2010), which provides insights on the sale of US patents. Several studies have contributed to this emerging stream of literature, most of them focusing on the market for US patents (e.g. Galasso et al., 2013; De Marco et al., 2017). Literature on European patents is scarcer (see Boesenberg and Egger, 2017 and Ciaramella et al., 2017 for notable exceptions).

Descriptive Statistics R&D Collaboration Figure 8.1 provides an overview of the proportion of patents co-applied between core European applicants and applicants from developing countries, by region (Panel A) and by technology field (Panel B).5 The number of co-applied patents grew from 16 in 1980 to 1,501 in 2010, leading to a compound annual growth rate (CAGR) of 16.3 per cent. To put that figure in perspective, collaboration between core European applicants and

4

5

The cost of registering a transfer of ownership at the EPO is EUR 100. There is no cost for registering a transfer of patent ownership at the DPMA since 2002. A8.1 explains how we allocate countries to groups of developed and developing countries. A8.3 discusses the allocation of patents to technological fields.

                        



Figure 8.1 R&D collaboration (applicant criterion), by region and technology Panel A. Patents co-applied between Core Europe and developing countries, by region in percentage. Panel B. Patents co-applied between Core Europe and developing countries, by technology field in percentage. Source: EPO PATSTATA database. Note: The figures above each bar indicate the total number of patents that were co-applied with developing countries.



          .

US applicants increased from about 500 patents in 1980 to more than 5,000 patents in 2010 and experienced a CAGR of 8 per cent. Thus, from a European perspective, collaboration with developing countries amounts to about one-third of the extent of collaboration with the United States. However, the growth rate has been more than twice as high. Regarding the regional breakdown, most of the growth in collaboration is driven by firms located in East Asia, notably China. Collaboration with Russia and other former Soviet Union countries is low and oscillates at around 100 patents since the 2000s. Finally, collaboration with Africa, Central and South America has slightly increased in total numbers, although it remains at low levels. As a proportion, patents arising from collaboration with non-Asian countries is shrinking fast – it represents a mere 17 per cent in recent years. Asia has now become the dominant collaborating partner among all developing regions. A strikingly similar pattern is obtained when looking at patents coapplied between US and Canadian applicants and applicants from developing countries (not shown), at least for recent years. The CAGR is higher than for Europe, reaching 24 per cent, and the breakdown in the 2010–2014 period is the following: East Asia (57 per cent); South Asia (33 per cent); Western Asia (